Browse the bounded catalog of registered Plasticity MCP operations or invoke one by its exact tool name. Use toolName='catalog' with query/offset/limit to read descriptions and JSON input schemas; set toolName to a returned operation name and pass its arguments object to invoke it. Only registered MCP operations are callable; the selected operation's original Zod schema and safety checks are applied. This is not JavaScript execution.

List Plasticity document windows exposed on loopback CDP.

Report MCP runtime and reachable Plasticity windows without changing a document.

Interpret a bounded fastening request such as 'крепится на 4 болта M5x10 с гайками' or 'печать винта M6x20 и ответной части'. It defaults to combined strength and geometry intent, parses thread metadata, quantity, selected ISO/DIN head and drive family, explicit joint phrases and fixed/adjustable/pivot intent, and returns one active question package plus compatible Plasticity geometry/strength tools. Its bounded catalog recognizes common hex, socket-cap, button, pan, countersunk and headless set-screw standards with source URLs; set-screw standards also identify flat, truncated-cone, dog, or cup points and require the mating contact function to be resolved. Printed screw, nut, and generic mating-part phrases route to the custom matched-thread tools without importing an ISO pitch. A slot is proposed only for explicit adjustment. It never treats nominal thread diameter as a finished hole, insert pocket, head recess, or nut envelope.

Persist an immutable physically tested rounded-print-v1 thread fit for one exact printer, material, slicer profile, nozzle, layer height, orientation, and thread definition. Requires explicit confirmation that a real specimen completed full-travel testing; geometric interference checks alone are not accepted. Repeating the same record is idempotent.

List immutable local physical thread-fit qualifications, optionally filtered by printer, material, slicer profile, rounded thread dimensions, handedness, or fit class. This registry is available without Workbench and contains no inferred or geometry-only passes.

Find a physical rounded-print-v1 thread-fit qualification for an exact process and thread definition. A record is eligible only when its tested engagement is at least the requested engagement. Conflicting qualified clearances return ambiguous and are never selected silently.

Check whether the nominal length in an ISO metric screw or bolt designation fits an explicit clamped stack and either a nut or threaded receiver. All layers, nut/washer envelope, engagement, tip clearance, and the product's under-head versus overall length datum remain explicit inputs. This deterministic check does not claim strength, preload, access, fit, or thread-stripping capacity.

Connect this MCP process to one explicit Plasticity window and return its compact initial scene summary. Body bounds/counts are paginated; use plasticity_body_info for exact topology of a selected body.

Read a compact document summary with identity, revision, exact body bounds, topology counts, and Undo/Redo state. Body summaries are paginated with bodyOffset/bodyLimit (default 0/50, maximum 200); follow bodyPagination.nextOffset and pass the prior page's revision as expectedRevision so scene edits cannot mix pages. Use plasticity_body_info for exact face/edge/vertex geometry of one body, or plasticity_list_bodies for paginated topology details.

Read bodies, linked instances, approximate reference meshes, groups, faces, edges, regions, and native Wire boundary/control handles currently selected by the user in Plasticity.

Replace the current Plasticity selection with stable body IDs.

Replace the current Plasticity selection with whole native Wire curves, using current revision-bound Wire IDs. The returned curveIds identify selected Wires; bodyIds may also contain the same native Wire IDs.

Replace the current Plasticity selection with a mixed set of current bodies, linked instances, approximate reference meshes, and native groups so the agent can point out assembly content in the application.

Replace the current Plasticity selection with current imported STL/OBJ reference meshes so the agent and user can point at the same approximate reference objects in the application.

Replace the current Plasticity selection with exact revision-bound B-Rep faces so the agent can point out surfaces in the application.

Replace the current Plasticity selection with exact revision-bound B-Rep edges so the agent can point out boundaries in the application.

Replace the current Plasticity selection with revision-bound Wire boundary vertices and interior B-Spline control points so the agent can point out the handles it will edit. References must come from plasticity_list_curve_control_points.

Rigidly rotate and move current bodies so the center and normal of one exact planar source face align with a fixed exact planar target face. The opposed relation seats outward face normals against each other; same keeps them parallel. Positive gap is measured from the target along its outward normal. The normal-to-normal rotation is the shortest rotation and does not independently align in-plane edges. All moving bodies preserve their relative placement and the alignment occupies one native Undo step.

Rigidly align exact native cylinder axes for one or more moving bodies in one Undo step. The preserve mode removes only transverse axis offset and keeps the cluster's axial position; anchor aligns the source axis origin to the target origin plus a signed axial offset along the target axis. The same/opposed relation controls axis direction, and an optional rotation around the fixed target axis controls roll. This is a direct placement, not a persistent concentric constraint.

Rigidly translate one or more moving bodies so one exact current B-Rep source vertex reaches a fixed target vertex plus an explicit world-space millimeter offset. The moving bodies retain their relative placement and the operation occupies one Undo step. This is direct placement, not a persistent coincident constraint.

Rigidly align one exact native Line edge from a moving body set to a fixed Line edge in one Undo step. Source and target edge midpoints are aligned with an optional signed axial offset along the fixed target tangent; same/opposed controls tangent direction, and rotationAroundAxisDeg controls roll around the fixed target line. Re-read the exact edges after placement because native edge parameter direction is topological, not a semantic assembly direction.

Check explicit pairs of current Solid bodies for exact volumetric interference by intersecting native B-Rep clones in a temporary database. The check does not change the document or Undo history. A no-volumetric-interference result does not distinguish touching from separation and is not a minimum-clearance measurement.

Measure exact native B-Rep volume, surface area, and volume centroid for one or more current Solid bodies without changing the document or Undo history. Totals include the volume-weighted centroid. This is geometric evidence; physical mass still requires an explicit qualified density.

Measure exact native B-Rep area, full trimmed boundary length, area centroid, surface type, and outer/inner loop counts for selected current Solid or Sheet faces without changing the document or Undo history. Totals include the area-weighted centroid; summed boundary length counts shared edges independently for each selected face.

Capture an in-memory scene baseline before manual edits. Returns only the snapshot ID, document/revision identity and object counts; use that ID with plasticity_changes_since or plasticity_wait_for_change.

Return a structured diff from a captured scene baseline. Exact changed-body B-Rep descriptors are summarized and paginated (bodyOffset default 0, bodyLimit default 20, maximum 100); follow bodyPagination.nextOffset with expectedRevision set to current.revision. Use plasticity_body_info for detailed topology of selected bodies. The diff also reports sketch Regions, construction planes, materials, visibility, instances, groups, and selection. Check sceneChanged for a reportable scene edit; revisionChanged can be true without a reportable scene diff.

Wait for a reportable scene edit, then return a compact paginated diff. Revision-only advances do not end the wait. Follow bodyPagination.nextOffset with expectedRevision set to current.revision; use plasticity_body_info for detailed topology of selected bodies.

List exact native B-Rep body details with topology IDs. Results are paginated (bodyOffset default 0, bodyLimit default 10, maximum 100); follow bodyPagination.nextOffset and pass the previous page's revision as expectedRevision so edits cannot mix pages. Prefer plasticity_body_info when only one body is needed. Cone faces also report native basis radius, axis origin/direction, and semi-angle in radians when Plasticity provides them.

List revision-bound planar regions generated by closed coplanar curves.

List exact revision-bound curve fragments created by native intersections.

List exact revision-bound endpoints of open native Wire bodies.

List every exact native Wire vertex with its body, position, endpoint flag, and adjacent native segment entity IDs. Use the returned revision-bound bodyId/vertexId pairs to select profile corners for curve filleting.

List exact start/end points and tangents for every native Wire segment.

Evaluate exact native positions and unit tangents at normalized parameters from 0 to 1 on current Wire segments. Parameters follow each segment's start-to-end direction from plasticity_list_curve_directions. This is read-only and is useful for placing or verifying local curve edits.

Read compact exact native B-Rep structure for selected current Wire bodies: segment type and length, analytic Circle center/radius/normal, plus NURBS degree, control-point count, Plasticity's raw span count, active normalized span count, carrier-knot parameters mapped into each segment's normalized coordinates, multiplicities, rationality, and periodicity when available. Periodic native span count can include wrapped extension knots; activeSpanCount counts only intervals across the normalized edge. A knot with withinSegment=false belongs to the underlying carrier outside that trimmed segment.

Read compact exact native B-Rep structure for selected current Solid or Sheet faces without changing the document: carrier surface type, whether the face is trimmed, face and natural UV parameter bounds, plus B-Surface degrees, span counts, control-point counts, and rationality. UV values are native parameters rather than millimeter distances.

Ask the native B-Rep kernel whether each selected current Wire is exactly planar and return its native plane origin and normal when available.

List every editable native control handle for selected current Wire bodies. Boundary vertices and interior B-Spline control points are returned separately with revision-bound references, positions in millimeters, and native local positive-U/negative-U unit slide directions. Handle positions and directions come from Plasticity's native editor representation; use exact B-Rep measurements to validate the resulting curve geometry.

List exact revision-bound native intersections between Wire bodies.

Read one body from the current document revision.

Read exact in-plane centers and diameters from two or more explicitly selected cylindrical faces on one current Solid. The returned revision-bound native B-Rep evidence and assignments can be passed into fastener-group load distribution; coaxial duplicate faces, stale references, and axes not normal to the supplied frame are rejected.

Measure and optionally check a fastener group on one exact rectangular planar face. Returns center-to-edge and hole-edge distances, every pair's center spacing and remaining ligament, plus clearance for supplied circular head, washer, nut, or driver envelopes. Optionally provide opposedFaceId to verify the matching perforated opposite face and exact native B-rep plate thickness. A pass is returned only when explicit layout requirements with a recorded basis are supplied. Geometry verification remains a layout check, not a strength or tool-motion analysis.

Find current B-Rep faces semantically by body, surface type, normal, radius, center, bounds, edge count, or adjacency.

Find current B-Rep edges semantically by body, curve type, direction, exact length, center, bounds, or adjacent faces.

Save a semantic face or edge query and re-evaluate it after topology changes.

Re-evaluate all semantic named selections against the current document revision.

Delete one MCP semantic named selection without changing Plasticity geometry.

Read the private local construction-event history, including operation inputs, outcomes and compact added/removed/modified B-Rep measurements. History survives MCP restarts and does not require a connected Plasticity window. Use offset/limit for paging; records are historical and never authorize replay.

Read current-process MCP mutation inputs and compact B-Rep change measurements in pages. This does not replay commands. The response also detects document changes made outside the journal and compares the live document against the last durable construction event.

List persistent CAD reference import provenance records (STEP, Parasolid, and approximate 3MF reference meshes) from the local MCP store. Records are historical evidence tied to the document/revision at import time; their body or mesh IDs are not current references. Use offset/limit to page results.

List persistent CAD reference import provenance records (STEP, Parasolid, and approximate 3MF reference meshes) from the local MCP store. Records are historical evidence tied to the document/revision at import time; their body or mesh IDs are not current references. Use offset/limit to page results.

Read one persistent CAD reference import provenance record (STEP, Parasolid, or 3MF). STEP and Parasolid records contain exact native B-Rep measurements; 3MF records contain approximate reference-mesh bounds and topology counts, not editable B-Rep. The record is historical; verify the current scene before using any stored body or mesh ID.

Read one persistent CAD reference import provenance record (STEP, Parasolid, or 3MF). STEP and Parasolid records contain exact native B-Rep measurements; 3MF records contain approximate reference-mesh bounds and topology counts, not editable B-Rep. The record is historical; verify the current scene before using any stored body or mesh ID.

Measure a body's axis-aligned native B-Rep bounds in millimeters.

Measure the exact straight-line distance and XYZ delta between two explicit coordinates or current native B-Rep vertices, edge midpoints, or face points. Topology references are revision-bound.

Measure the exact distance from an explicit coordinate or current B-Rep vertex to a finite straight B-Rep edge centerline. Returns the projected point, the clamped closest point on the finite edge, supporting-line distance and finite-segment distance. This is centerline geometry, not minimum clearance to the owning faces or bodies; references are revision-bound.

Estimate point distance to a non-linear, non-circular Solid/Sheet B-Rep edge or native Wire segment using an adaptively sampled polyline of native Plasticity edge evaluations. For Solid/Sheet, pass bodyId plus edgeId; for Wire, pass bodyId plus segmentEntityId from plasticity_list_curve_directions. Returns an explicitly approximate closest point/distance, sample count and maximum midpoint chord deviation observed at those samples. toleranceObserved only means the sampled deviation criterion was met; it is not a certified global error bound. For dimensions that need exact 0.01 mm proof, use exact line/circle tools or request a more suitable native analytic method; never report this estimate as exact or as body clearance. Current body/topology references are revision-bound.

Measure exact distance from an explicit coordinate or current B-Rep vertex to a native circular boundary. Use bodyId plus edgeId for Solid/Sheet topology, or bodyId plus segmentEntityId from plasticity_list_curve_directions for a Wire. Returns distance to the supporting circle, minimum distance to the trimmed arc, closest point, normalized arc parameter, and endpoint-clamp status. Requires verified native circle basis and trim samples; unsupported or stale references are rejected. This is distance to the edge curve centerline, not clearance to adjacent faces or the owning body.

Measure the exact minimum distance from an explicit coordinate or current B-Rep vertex to a trimmed planar face. Supports closed polygonal loops, complete circles, and exact trimmed circular arcs, including holes; returns signed supporting-plane distance, minimum distance to the actual trimmed face, closest point, and whether that point lies in the face interior or on a boundary edge. Other curved boundaries, non-planar faces, incomplete topology, and stale references are rejected; this does not measure clearance between two bodies.

Measure the angle between two current planar B-Rep faces and, when parallel, the signed/absolute separation of their infinite supporting planes in millimeters (separationKind='supporting-planes'). It does not test overlap of the trimmed faces or measure their actual gap/clearance; use only when supporting-plane spacing is the intended quantity.

Measure exact minimum distance between two revision-bound parallel planar B-Rep faces, including polygonal regions, complete circular boundaries and exact trimmed circular arcs (including holes). Computes planar overlap or boundary separation together with the normal plane gap and closest 3D points. Requires parallel faces within 1e-7 degrees and closed boundaries composed of straight edges and exact circular edges, at most 512 edges per face; other curved edges, nonparallel faces, incomplete topology and stale references are rejected rather than approximated. This is face-to-face clearance, not body-to-body collision detection.

Measure exact minimum Euclidean distance between two nonparallel trimmed planar B-Rep faces with simple straight-edged polygon boundaries. Concave outlines, holes and multiple nested/disjoint loops are supported; face regions use even-odd loop filling. Reports exact closest 3D points and zero when the finite face regions intersect. Limits each face to 512 boundary vertices and 4096 exact decomposition triangles. Faces with parallel supporting planes, circular/curved edges, self-intersecting or touching loops, incomplete native topology, or stale references are rejected; use plasticity_measure_parallel_planar_face_clearance for supported circular boundaries. This measures a selected face pair, not whole-body clearance or collision.

Measure the exact thickness of 1-64 clamped Solid layers from explicit pairs of parallel native planar faces aligned with the fastener axis. Returns revision-bound gripItems that can be passed directly to plasticity_check_fastener_stack. Washers or other layers must be modeled and selected explicitly.

Measure the exact angle and closest approach between two current linear B-Rep edges. Returns both infinite supporting-line clearance and the exact distance/closest points between the finite edge centerline segments. Finite-edge distance is centerline geometry only; it is not minimum clearance between the owning faces or bodies.

Sample Plasticity's native B-Rep curvature vector at 100 positions along each selected current edge. Reference Wire segments by bodyId plus segmentEntityId from plasticity_list_curve_directions; reference Solid or Sheet edges by bodyId plus edgeId from state. Returns curvature magnitude in 1/mm, finite radii of curvature in millimeters, extrema locations, and whether any sampled point is straight. This compact sampled analysis does not use the display mesh, does not claim an exact continuous maximum, and does not change the document.

Sample exact native B-Rep face normals on a finite interior grid and classify each selected Solid or Sheet face as positive, negative, neutral, or mixed draft relative to an explicit pull direction and minimum draft angle. Signed draft is asin(normal dot pull): walls parallel to pull are 0 degrees, outward normals toward pull are positive, and opposite normals are negative. The compact extrema do not prove continuous extrema, mold parting, release paths, or print support requirements.

Sample Plasticity's native B-Rep continuity evaluator at 100 positions along each selected current Solid or Sheet edge shared by exactly two faces. Returns maximum G0 position deviation in millimeters, G1 normal-angle deviation in degrees, Plasticity's dimensionless relative G2 curvature deviation, their locations, and hierarchical G0/G1/G2 tolerance results. This is native surface sampling rather than an exact continuous maximum; it does not use the display mesh and does not change the document.

List native measurements stored in the Plasticity document with stable measurement IDs, current topology targets, and exact values when both targets still resolve.

Create a persistent native Plasticity point-to-point distance measurement between two current B-Rep vertices. The measurement participates in document Undo/Redo and remains attached to its topology.

Create a persistent native Plasticity point-to-point distance measurement between two current Solid or Sheet topology points. Each endpoint may be an exact vertex, edge midpoint, or face center. The measurement remains attached to topology, participates in Undo/Redo, and is listed with stable measurement and topology identities.

Create a persistent native Plasticity radius measurement on one exact current circular edge. Reference a Wire segment by bodyId plus segmentEntityId from plasticity_list_curve_directions, or a Solid/Sheet edge by bodyId plus edgeId from state. The measurement remains attached to topology, participates in Undo/Redo, and plasticity_list_measurements reports both radius and derived diameter in millimeters.

Delete one current native Plasticity measurement by stable measurement ID through document Undo/Redo history.

List the active native Plasticity viewport section plane, including its session-stable analysis ID, origin, clipped-half-space normal, visibility, and number of viewports using it. This reads the section currently applied by Plasticity shading state; sections are viewport state and do not modify B-Rep or Undo history.

Apply a native Plasticity section plane to the current viewport from a world-space origin and normal. The normal points toward the half-space to clip. An optional nonparallel xDirection sets the plane helper orientation; otherwise MCP derives one. Plasticity supports one active section plane at a time; delete it before applying another. The change is immediately visible in Plasticity and changes the MCP revision, but does not modify B-Rep or Undo history.

Clear the active native Plasticity section plane by its current session-stable analysis ID. The viewport returns to an unclipped state; B-Rep and Undo history are unchanged, while the MCP revision is updated.

List current native linked instances, their revision-bound IDs, source bodies, and exact transforms. Instance geometry remains linked to its source until realized.

Create one native linked instance from a current source body, optionally translated in millimeters. Later source-body edits propagate through the instance.

Create independent exact native copies of current Solid or Sheet bodies, preserve every source, and translate the complete copied set by an explicit world-space millimeter delta in one Plasticity history step. The copies receive new stable body IDs and are ordinary editable B-Rep bodies rather than linked instances. A zero translation deliberately creates coincident copies.

Move current native linked instances by a world-space millimeter delta without changing their source geometry.

Rotate current native linked instances around a world-space pivot and axis. The angle is in degrees.

Scale current native linked instances around a world-space pivot with positive XYZ factors, preserving their link to the source geometry.

Convert current linked instances into independent native Plasticity bodies. Use this before an instance needs geometry edits, Boolean operations, or independent fabrication changes.

Delete current native linked instances through Plasticity Undo/Redo history without deleting their source bodies.

List imported STL/OBJ reference meshes with stable IDs, source paths, approximate world-space bounds, buffer counts, and transforms. Reference-mesh bounds come from tessellated data and are never native B-Rep dimensional proof.

Move current approximate STL/OBJ reference meshes by a world-space millimeter delta in one Plasticity history step.

Rotate current approximate STL/OBJ reference meshes around an explicit world-space pivot and nonzero axis. The angle is in degrees and the edit occupies one Plasticity history step.

Scale current approximate STL/OBJ reference meshes around an explicit world-space pivot with positive XYZ factors in one Plasticity history step. Re-read the returned mesh bounds; scaling does not make the source dimensionally authoritative.

Delete current approximate STL/OBJ reference meshes through Plasticity Undo/Redo history without touching native B-Rep bodies.

Rename one current imported STL/OBJ reference mesh through Plasticity document history.

List the current native Plasticity group hierarchy, active group, direct body, linked-instance, and reference-mesh members, visibility, and lock state.

Create one native Plasticity group around the supplied current bodies, linked instances, approximate reference meshes, or child groups. The root Scene group cannot be nested.

Move current bodies, linked instances, approximate reference meshes, or whole child groups into an existing destination group in one native history step. Group cycles are rejected.

Rename one current non-root native Plasticity group through document history.

Make a current native Plasticity group the active destination for newly created objects. Use group ID 0 to return creation to the root Scene group.

Dissolve current non-root Plasticity groups and promote their contents to each parent without deleting the contained geometry.

Set exact native visibility for current bodies, linked instances, approximate reference meshes, or groups. Hidden geometry remains in the document and can be restored with the same tool.

Set the native Plasticity lock state for current bodies, linked instances, approximate reference meshes, or groups. Locked geometry remains readable but resists manual selection and editing.

Run Plasticity's native B-Rep Check and report exact topology closure and solid printability for current revision-bound bodies.

Page through native renderer bindings found in Plasticity 26.1.3. Defaults to 100 bindings per page; filter by case-insensitive substring with query. Returns total and matching counts plus nextOffset so large binding lists do not flood MCP context.

Read the document after a timeout or disconnect and allow later mutations without retrying the uncertain command.

List Plasticity document appearance materials and each body's assigned material ID. These visual properties are not manufacturing or strength evidence.

Define a revision-bound point from coordinates or exact current topology.

Define a revision-bound axis from points, coordinates, a linear edge, or a cylindrical face.

Create a native saved construction plane from an exact plane definition.

List session datums and current standard and saved construction planes.

Activate a current construction plane in the selected Plasticity window.

Remove one current saved construction plane. Standard planes are protected.

Re-resolve a geometry-backed datum after a document revision change.

Create an axis-aligned exact CAD box. Coordinates and size are millimeters.

Create an exact CAD cylinder along a world-space axis. Dimensions are millimeters.

Create an exact cone or conical frustum Solid and preserve its editable meridional polyline profile in one Plasticity history step. The bottom center, unequal bottom and top radii, height, axis, and nonparallel radial direction are explicit. Use topRadiusMm=0 for a pointed cone and plasticity_create_cylinder when the radii are equal.

Create an exact ring torus Solid and preserve its editable native circular profile in one Plasticity history step. Center, major radius to the tube centerline, minor tube radius, symmetry axis, and radial zero direction are explicit; the major radius must exceed the minor radius.

Cut an exact through hole with a concentric conical countersink into one Solid. The caller supplies the finished through diameter, major diameter, included angle, material depth, and an in-plane radial direction from the selected standard or manufacturer record. The recipe preserves its editable annular meridional Wire and records through cutter, profile, Revolve, and Boolean as four native history steps.

Cut 2-64 equal through holes with concentric conical countersinks at explicit entry centers on one current Solid. Each center creates a through cutter and an editable meridional Wire revolved into a countersink cutter; one final Boolean consumes every cutter while preserving the source Wires. Finished dimensions must come from the selected fastener standard, fit, process, and actual material depth.

Cut an exact blind regular-hex pocket into one Solid. Across-flats size, pocket depth, material depth, orientation, and clearance must come from the selected nut, manufacturing process, and access requirements; nominal thread diameter is not used as pocket geometry. The profile Wire remains editable and the recipe records profile, Extrude, and Boolean as three native history steps.

Cut 2-128 equal blind regular-hex nut pockets at explicit entry centers on one current Solid. Each center creates one editable profile and one extruded cutter; one final Boolean consumes every cutter while preserving the source Wires. Across-flats size, depth, orientation, clearance, and material depth must come from the selected nut and manufacturing process.

Create a standalone exact one-start rounded thread intended for a matched printed part. It creates and preserves a native Helix, forms the ridge with Pipe and Boolean, and clips the result to the explicit crest diameter and length. Nominal diameter is geometry only: this rounded-print profile is not an ISO metric thread and needs printer/material/profile-qualified depth and pitch.

Cut an exact one-start rounded internal thread into one current Solid for a matched printed external thread. The bore, helical groove, normal profile clearance, material depth and overshoot are explicit. This custom rounded-print profile is not an ISO tap and must not be assumed compatible from an M designation alone.

Create an exact extruded hex nut with a matched one-start rounded internal print thread. Across-flats size, thickness, minimum remaining wall, pitch, depth and normal profile clearance are explicit; the source hex Wire and Helix remain editable. It is a custom printed mating part, not an ISO nut unless a separate verified standard-profile workflow is used.

Create an exact wrenchable hex-head screw with a standalone one-start rounded print thread. The thread is clipped to its explicit crest envelope and joined to an overlapping hex head; the source head Wire and Helix remain editable. This is a custom matched print profile and is not ISO metric hardware solely because its crest diameter is named M5 or similar.

Create a complete matched printable hex screw and nut from a designation such as 'printed M5x10 pair'. The designation supplies only crest diameter and screw thread length; pitch, rounded-profile depth, normal clearance, head and nut envelopes remain explicit and shared. Printer, material, slicer profile, nozzle, layer height, orientation, clearance evidence and sizing basis are recorded. An optional immutable physical qualification ID is checked against the exact process, thread definition, clearance, and required engagement before mutation. Without it, the result remains marked as requiring a physical fit test. This custom rounded-print-v1 pair is not ISO metric hardware.

Create one custom rounded-print-v1 hex screw and 2-8 separate hex nuts with explicit, unique normal profile clearances. Use this process-specific clearance ladder before choosing the working fit for a printer, material, slicing profile, and orientation. The result remains unqualified until the specimens are physically printed and tested; it is not ISO metric hardware even when the crest diameter is written like M5.

Cut an exact straight through-slot with semicircular ends into one Solid. Overall length includes the round ends and must exceed width; entry center, in-plane slot direction, cutting axis, material depth, positive edge distance, and clearance-adjusted finished dimensions are explicit. Exact tangency to an exterior boundary is invalid. The editable center profile remains and five native history steps are returned.

Cut 2-64 equal exact straight through-slots with semicircular ends at explicit entry centers on one Solid. Shared in-plane adjustment direction, overall length, finished width, cutting axis, actual material depth, and positive edge distance are explicit; exact tangency to an exterior boundary is invalid. One editable center profile and three cutters are created per slot; one final Boolean yields 4N+1 confirmed history steps.

Cut an exact through hole with a concentric flat-bottom counterbore into one Solid. The entry point lies on the target surface and the axis points into it. The recipe uses three explicit native history steps and returns each confirmed revision.

Cut 2-128 equal through holes with concentric flat-bottom counterbores at explicit entry centers on one current Solid. Each center creates one through cutter and one recess cutter; one final Boolean consumes every cutter. Finished dimensions must come from the selected head, fit, manufacturing process, and actual material depth.

Cut one exact round through-hole into a current Solid. The entry center lies on the target surface, the axis points into the material, and the finished diameter and material depth must come from the selected fit or qualified process rather than the nominal thread diameter. The cutter and Boolean are two explicit native history steps.

Cut 2-256 equal exact round through-holes at explicit entry centers on one current Solid. The shared axis points into the material; finished diameter and material depth must come from the selected fit or qualified process. Each native cylinder is a confirmed history step and one final Boolean consumes every cutter.

Cut one exact flat-bottom blind round hole into a current Solid. The entry center lies on the target surface, the axis points into the material, and the explicit finished diameter and depth must come from the selected tap, screw, insert, or qualified process rather than the nominal thread diameter or fastener length. Material depth is required and must exceed hole depth. The cutter and Boolean are two explicit native history steps.

Cut 2-256 equal exact flat-bottom blind round holes at explicit entry centers on one current Solid. The shared axis points into the material; qualified finished diameter and hole depth are explicit, and material depth must be greater than hole depth. Each native cylinder is a confirmed history step and one final Boolean consumes every cutter.

Cut an exact three-stage pocket for a heat-set insert: a deep pilot, insert bore, and wider shallow lead-in. The entry point lies on the target surface and the axis points into it. Explicit material depth must exceed pilot depth. The recipe uses four explicit native history steps and returns each confirmed revision.

Cut 2-64 equal exact three-stage heat-set-insert pockets at explicit entry centers on one Solid. Qualified pilot, insert-bore, lead-in, and material depths are shared; three native cylinders per center and one final Boolean yield 3N+1 confirmed history steps.

Create an exact cylindrical screw boss joined to an existing Solid, then cut a blind pilot hole from its top. The base point lies on the support surface and the axis points outward. The recipe uses four explicit native history steps and returns each confirmed revision.

Create 2-64 equal exact cylindrical screw bosses at explicit base centers, join all bosses to one existing Solid in one native union, then cut every blind pilot in one native difference. The shared axis points outward; base overlap gives every boss real union volume. The recipe returns 2N+2 confirmed history steps.

Create an exact rib from a closed coplanar world-space profile, extrude it by a signed thickness, and join it to one existing Solid. The profile Wire remains editable. The recipe uses three explicit native history steps and returns each confirmed revision.

Cut an exact rectangular array of round through-vents into one Solid. The first center lies on the entry surface, the axis points into it, and both array directions lie in that surface plane. One seed cylinder, one native rectangular pattern, and one Boolean produce up to 400 holes in three history steps.

Create an exact cantilever snap-fit beam with an integral end hook and join it to one Solid. The base point lies on the support, beam and thickness directions define the profile plane, and width is centered across its normal. The editable profile remains in the document and the recipe uses three native history steps.

Create an exact hollow cylindrical hinge barrel along a world-space axis and join it to one existing Solid. The bore diameter is the finished pin-clearance diameter. The recipe uses four native history steps and returns every confirmed revision.

Cut one or more exact circular cable channels along current editable Wire paths. The finished channel diameter includes required cable clearance. Native solid Pipe cutters are consumed by one Boolean while source Wires remain in the document. The recipe uses two history steps.

Cut an exact rectangular or rounded-rectangular connector opening into one Solid. The entry center lies on the target surface, the axis points into it, and width direction lies in that surface. The editable profile remains in the document; native extrusion, optional fillet, and Boolean steps are recorded separately.

Add an exact male lip and clearance-matched female groove to two existing axis-aligned enclosure halves. The seam origin is the lower-left outer corner at the mating plane. All temporary ring solids are consumed and eight native history steps are returned.

Add one exact cylindrical locating pin to a male Solid and cut its clearance-matched socket into a different female Solid. The base center lies on the mating plane and the axis points from the pin into the socket. Radial and axial clearances are explicit millimeter inputs.

Add 2-64 explicit cylindrical locating pins to a male Solid and cut matching radially and axially clearanced sockets into a different female Solid. Centers lie on the mating plane; the axis points from pins into sockets. One native union joins all pins and one native Boolean cuts all sockets. Clearances are exact geometric inputs, not process-qualified print-fit recommendations.

Build a split-half fastener pattern in one logical recipe: cut matched clearance holes through the male half, then create exact pilot, heat-set insert, and lead-in pockets from the mating plane into the female half. Supply paired coaxial centers, a resolved headed screw designation and under-head length, exact insert part number and HTTPS source, matching insert thread diameter/pitch, explicit engagement limits, and qualified process-specific pocket dimensions. The tool rejects thread or length mismatches before CAD mutation and verifies that all stations align within 0.01 mm. It does not infer insert geometry, thread capacity, print clearance, head seat, or joint strength.

Add an exact rectangular tongue to one Solid and cut its clearance-matched blind groove into another. The base center lies on the mating plane, the axis points from tongue to groove, and width direction lies in that plane. Both source profiles remain editable.

Create an exact flared trapezoidal male dovetail on one Solid and cut the radially and axially clearance-expanded female socket into another. rootWidthMm is the narrow root, each side flares by flareMm toward the tip, widthDirection lies in the mating plane, and the explicitly supplied clearances are geometric inputs rather than qualified print-fit recommendations. Both editable profiles remain in the document; six native history steps are returned.

Create an exact CAD sphere in millimeters.

Create a native polyline or closed profile in world coordinates or a referenced construction plane.

Create exact closed constant-width slot or channel profiles around one or more current planar Wire spines in one native Plasticity history step. Each source Wire is preserved and each result is a separate editable Wire suitable for Region extrusion or cutting. Width is the full finished profile width. A lone straight segment does not define a unique native plane; use a planar multi-segment or curved spine, or use plasticity_create_slotted_hole for a straight fastener slot.

Create a native interpolating NURBS curve through three or more world-space points in millimeters.

Create a constant-radius native helix around a world-space axis. Coordinates and radius are millimeters.

Create a native circle in world coordinates or a referenced construction plane.

Create one exact native circle from the two endpoints of its diameter. In world coordinates, the explicit plane normal must be perpendicular to the diameter; plane-local inputs inherit the referenced construction plane. The result is a closed Wire with an automatic Region.

Create one exact native circle through three distinct non-collinear points. The three points determine its plane, center, and radius; plane-local inputs are resolved through the referenced construction plane. The result is a closed Wire with an automatic Region.

Create one exact native circular arc from center, radius, start angle, and signed sweep. Positive sweep is counterclockwise around the plane normal; the sweep magnitude must be below 360 degrees. Inputs may use world coordinates or a referenced construction plane.

Create one exact native circular arc from a start point through a second point to an end point. Point order selects the minor or major arc and preserves start-to-end curve direction. Inputs may use world coordinates or a referenced construction plane.

Create one exact native circular arc tangent to a selected current Wire segment at its start or end and terminating at an explicit point. Use plasticity_list_curve_directions to obtain the segment entity and its exact endpoints. flipTangent selects the opposite tangent sense, which can select the major rather than minor arc. The end may use world coordinates or a referenced construction plane.

Create one exact fixed-radius native circle tangent to two current Wire segments. Use plasticity_list_curve_directions immediately before this call. solutionPointMm selects the intended solution near one of the possible circle centers; it is not an additional geometric constraint. World-space inputs require the sketch-plane normal, while plane-local inputs inherit the referenced construction plane. Both source Wires remain unchanged.

Create one independent native B-Spline between two exact current Wire segment endpoints while preserving both source Wires. Use plasticity_list_curve_directions immediately before this call. G0 matches position, G1 adds tangent continuity, G2 adds curvature continuity, and G3 adds third-order geometric continuity at each end. The operation occupies one Plasticity history step.

Create one independent native B-Spline between two exact current open Wire vertices while preserving both source Wires. Use plasticity_list_curve_vertices immediately before this call and pass its numeric vertexId values. G0 through G3 continuity is selected independently at the two ends; internal or stale vertices are rejected before mutation.

Create one independent native B-Spline between explicitly selected endpoints of two exact current Solid or Sheet edges while preserving both source bodies. Each reference must pair a current edgeId with one of that edge's returned vertexIds. G0 through G3 continuity is selected independently at the two ends; exact endpoint and tangent read-back is required after the one-step operation.

Create one exact native closed elliptical Wire and Region from explicit major and minor radii. The major axis follows xDirection rotated by angleDegrees in the selected plane. Inputs may use world coordinates or a referenced construction plane.

Create one exact native closed regular-polygon Wire and Region with 3-256 vertices. radiusMode=circumradius places every vertex on the supplied radius; radiusMode=inradius uses the supplied center-to-edge distance. Inputs may use world coordinates or a referenced construction plane.

Create an exact native rectangular Wire and Region from its center, width, and height in world coordinates or a referenced construction plane. Positive angle rotates the width axis around the plane normal in degrees.

Create Plasticity-native closed Wire outlines for text at a world or construction-plane baseline origin. Font size is nominal millimeters; inspect the resulting exact curve bounds before using the outlines for embossing, engraving, or clearance-critical geometry. Plasticity may create multiple Wires and Regions.

Move bodies. Requires the current revision; delta is millimeters.

Rotate bodies around a world pivot and axis. Angle is degrees.

Apply a Workbench workbench_assess_printability rotationDeg to one or more exact native Solids as one rigid group in one Plasticity history step, around the union B-Rep bounding-box center. Pass the same expectedSizeMm from that DFM result; the tool refuses a rotation whose predicted group bounds disagree with it, then returns actual native group bounds and verification status. The orientation is rotation about X, then Y, then Z (Workbench Euler XYZ convention); it does not split the bodies or prove slicer fit, so re-run DFM and slicing afterward.

Scale bodies around a world pivot with positive XYZ factors.

Set the exact local width, length, and height of one current Solid that Plasticity still recognizes as a dimensionable block. This is a one-step native direct edit centered on the existing block; it does not create a persistent parametric constraint. Read the resulting B-Rep dimensions back after the edit.

Set the exact radius of one current cylindrical B-Rep face through Plasticity's native direct-dimension command. This changes recognized coaxial geometry in one Undo step; it is not a fillet command or persistent parametric constraint. Read the resulting cylindrical face radius back after the edit.

Set Plasticity's exact local width and length for one closed planar Wire that the native dimension command recognizes as a rectangle. The profile remains centered and its Region updates in one Undo step. This is a direct edit, not a persistent constraint; read the resulting Wire B-Rep bounds back.

Apply exact CAD union, difference, or intersection.

Split current Solid or Sheet bodies with current planar cutter faces. Cutter bodies are preserved. To trim an open Sheet, inspect the returned exact Sheet parts and delete only the unwanted part at the returned revision.

Split one current Solid into exactly two native Solid parts with one explicit plane that crosses its interior. The tool derives an oversized planar cutter from exact B-Rep bounds, cuts the Solid, verifies both native volumes sum to the original within tolerance, checks unrelated bodies are unchanged, and removes its temporary cutter geometry. It does not add an assembly joint; create a qualified locating-pin or tongue-and-groove joint afterward if required. The plane origin, normal, and in-plane x direction are millimeters/world vectors and the tool occupies four native history steps.

Split one current Solid into a grid using an ordered list of explicit world-space planes. Each plane is applied only to current Solid parts whose exact B-Rep bounds it crosses; each native cut must produce exactly two valid Solid results and preserve volume, temporary cutters are removed, and the final part volumes are checked against the source. Planes are processed sequentially and partial results remain if a later cut fails; reconcile and inspect before continuing. This tool does not choose planes from a Workbench splitPlan, orient the source, or add joints. Each successful cut uses four native history steps.

After orienting a current Solid to the printer's world X/Y/Z axes, split it into an even grid sized from its exact native B-Rep bounds and the selected profile's usable build volume in millimeters. The tool derives world-space cut planes, performs the bounded native multi-plane recipe, validates exact total volume and native Solids, and verifies every result's exact bounds fit the usable volume within 0.01 mm. It does not rotate the part or create assembly joints; a failed operation can leave confirmed earlier cuts, so inspect the revision and scene before continuing.

Assign an existing Plasticity appearance material, create and assign a bounded color/roughness/metalness/opacity appearance, or clear an assignment with materialId 0. This changes display appearance only and uses one native Undo step.

Fillet current edge IDs with a radius in millimeters.

Chamfer current edge IDs by an equal distance in millimeters.

Remove every native fillet recognized on one or more current Solid or Sheet bodies.

Change recognized native fillet faces by a signed radius delta in millimeters.

Extrude current face IDs by a distance in millimeters.

Extrude one unambiguous closed planar Wire profile into a native Solid.

Extrude one or more explicit revision-bound planar regions into native solids.

Create native offsets from planar Wire bodies. Signed distance follows each Wire orientation and is measured in millimeters.

Create one or two native signed offsets from explicit Regions in the same sketch, preserving the source curves.

Remove one or more explicit revision-bound curve fragments using Plasticity's native Trim operation.

Extend one or more explicit revision-bound open Wire endpoints by a positive distance in millimeters.

Convert one or more exact current interior or closed Wire vertices into native B-Spline control vertices in one Plasticity history step. References must come from plasticity_list_curve_vertices at the current revision. Open endpoints are not convertible. The Wire path, segment structure, length, and topology change, so discard every old vertex and segment reference and inspect the returned Wire before continuing.

Round one or more exact interior or closed Wire vertices with Plasticity's native curve fillet in one history step. Vertex references come from plasticity_list_curve_vertices and are revision-bound; the positive radius is in millimeters. The edited Wire may receive a new stable body ID, so use the returned state before further work.

Split compound native Wire bodies into separate editable curve bodies.

Create independent exact native copies of one or more current Wire bodies in place while preserving the sources. The copies receive new stable body IDs and can be transformed or edited separately.

Create independent exact native Wire copies of the boundaries of explicit current planar Regions while preserving the source Wire geometry. Coincident boundary copies make Plasticity recompute automatic Regions, so all previous Region references become stale; use the returned state or plasticity_list_regions before downstream work.

Join two or more native Wire bodies into one editable compound curve.

Rebuild current native Wire bodies in one Plasticity history step. tolerance asks Plasticity to fit within an explicit millimeter target; control-points sets the native control-point count; degree-spans sets exact NURBS degree and span count. Read plasticity_inspect_curve_structure before and after, and measure geometric drift because a requested fit setting is not independent proof of deviation.

Raise the native degree of every B-Spline segment in one or more current Wire bodies by one while preserving its shape. This adds edit freedom without adding shape detail. Inspect native curve structure before and after; each call occupies one Plasticity history step.

Insert native knots into one or more current B-Spline Wire bodies while preserving degree and shape. This adds local edit points without changing the curve's path. Inspect native curve structure before and after; each call occupies one Plasticity history step.

Insert one native knot into one exact current B-Spline Wire segment while preserving its path and degree. Use plasticity_list_curve_directions immediately before this call and pass a normalizedParameter strictly between 0 and 1, where 0 is the returned segment start and 1 is its end. This adds one local control point in one Plasticity history step; inspect native structure and functional geometry afterward.

Split one exact nonperiodic current Wire segment into two consecutive segments at a normalized parameter strictly between 0 and 1. Evaluate the intended point first with plasticity_evaluate_curve_segments. The Wire body remains one body and its path is preserved, but every old segment reference becomes stale. Full periodic circles are rejected because one split point only relocates their seam.

Orthogonally project one or more current Wire bodies onto an explicit world-space plane. This changes a spatial curve's path and may reverse its parameter direction, unlike degree elevation or subdivision. Verify planarity, endpoints, direction, and functional measurements afterward. The operation occupies one Plasticity history step.

Move one or more exact current Wire control handles by a shared world-space millimeter delta in one Plasticity history step. References must come from plasticity_list_curve_control_points at the current revision. Boundary vertices edit curve ends or joins; interior control points reshape B-Splines. Re-read control points and exact B-Rep geometry afterward.

Slide one or more exact current Wire handles by a positive millimeter distance along each handle's local positive-U or negative-U control-polygon direction. Use the unit directions returned by plasticity_list_curve_control_points to predict the world-space result. All handles share one direction sense and distance, and the edit occupies one Plasticity history step. Re-read handles and exact B-Rep geometry afterward.

Rotate exact current Wire control handles around an explicit world-space pivot and axis in one Plasticity history step. References must come from plasticity_list_curve_control_points at the current revision. The angle is in degrees; re-read handles and exact B-Rep geometry afterward.

Scale exact current Wire control handles around an explicit world-space pivot with positive XYZ factors in one Plasticity history step. References must come from plasticity_list_curve_control_points at the current revision. Re-read handles and exact B-Rep geometry afterward.

Delete one or more current interior B-Spline control points from a single Wire in one Plasticity history step. References must come from plasticity_list_curve_control_points at the current revision. This changes the curve path and reindexes the remaining control-point IDs, so discard every old handle reference and re-read structure, handles, endpoints, tangents, and functional geometry afterward.

Reverse the native direction of one or more current Wire bodies.

Reverse the native surface normal orientation of one or more current Sheet bodies.

Create exact native Wire silhouettes from current Solid or Sheet bodies on an explicit current construction plane. Source placement keeps each outline at the source silhouette plane; workplane placement projects it onto the selected plane. Source bodies are preserved and the selected plane becomes active.

Project native Wire bodies onto a surface body along an explicit world-space vector while preserving the source curves.

Create independent exact Wire curves at every native intersection between one Solid or Sheet target and one or more Solid or Sheet tools. All source bodies are preserved and the complete operation occupies one Plasticity history step.

Create an independent 3D Wire by intersecting the bidirectional extrusion surfaces of two distinct native Wire bodies. Supply one explicit world-space projection direction for each source and a depth in millimeters large enough for both temporary surfaces to overlap. Source curves are preserved and the operation occupies one Plasticity history step.

Insert native U- or V-isoparametric edges into one current Solid or Sheet face. This splits the selected face in place while preserving the body ID and occupies one Plasticity history step; it does not create independent Wire bodies. U/V follow the face's native parameterization, so inspect the resulting analytic surfaces, dimensions, topology, and mass properties instead of assuming a world direction or unchanged numerical integration.

Raise the U and V degree of every selected current native B-Surface face by one Plasticity step in one history entry. Plasticity 26.1.3 may also change span/control-point counts and exact geometry; inspect the surface structure, bounds, topology, and functional dimensions before and after instead of assuming shape preservation.

Restore selected current Solid or Sheet faces to the natural bounds of their carrier surfaces in one native history entry. The selected trim boundaries are discarded, the result can overlap neighboring geometry, and all topology references become stale; inspect surface structure, bounds, validation, and intersections before continuing.

Project native Wire bodies onto a Solid or Sheet and split its exact faces along the projected curves.

Split a Solid or Sheet target along exact intersections with preserved Solid or Sheet tool bodies.

Sweep one or more explicit closed Regions along a native Wire spine to create exact Solid geometry.

Loft an ordered list of closed Regions from different sketch planes into exact capped geometry, optionally shaped by native Wire guide curves that intersect every profile.

Create one independent native loft surface through an ordered list of current Wire profiles while preserving every profile and guide. Optional Wire guides must intersect every profile. Closed mode requires at least three profiles and closes the loft sequence; natural, unconstrained, or clamped native curvature plus positive dimensionless end magnitudes control shape. The operation never joins the result to source bodies; verify the returned Sheet topology and bounds.

Create one independent native capped loft Solid through an ordered list of exact planar faces from different current Solid or Sheet source bodies while preserving every source. Optional current Wire guides must intersect every profile. Natural, unconstrained, or clamped native end conditions and positive magnitudes control the two ends; verify the returned B-Rep rather than treating magnitude as a millimeter distance.

Create native Sheet bodies that fill explicit revision-bound closed Regions.

Create independent native Sheet patches from current closed Wire bodies while preserving every source Wire. Unlike planar Region patching, this accepts nonplanar closed boundaries and can create B-Surfaces. Re-read exact boundaries, surface structure, area, and native validity; the native fill is not a constrained engineering surface unless those properties are separately verified.

Create a native G2 transition surface between boundary sides of two Sheet faces. Pick points and width are millimeters; each pick point must lie on the intended boundary edge.

Sew two or more native Sheet bodies along coincident edges.

Create a native B-Surface constrained by paired 3D points and normal vectors.

Refit one exact current Solid or Sheet face as a native B-Surface with projected boundary edges. The positive millimeter tolerance is an approximation setting, not a measured deviation bound. Plasticity rebuilds the owning body in one history step and invalidates every prior topology reference; re-read dimensions, surface structure, continuity, mass properties, and native validity afterward.

Replace one or more exact current Solid or Sheet face surfaces with the carrier surface of one separate current replacement face. Plasticity extends or trims adjacent faces to meet the replacement surface and performs the direct edit in one history step. An external replacement body is preserved; the edited bodies keep their stable IDs, but all prior topology references become stale. Verify bounds, analytic surface type, mass properties, interference, and native validity afterward.

Copy exact current faces into separate native Sheet bodies while preserving their source bodies.

Create an exact planar native Sheet development from one current analytic Cylinder face while preserving the source body. This is geometric surface unwrapping; it does not add sheet thickness, bend radii, bend allowances, or manufacturing compensation. Plasticity chooses the seam and planar placement, so inspect the returned Sheet bounds and edges.

Calculate an area-preserving annular-sector profile from one complete native conical-frustum face bounded by two full circles and one straight seam. Returns exact B-Rep source identity, radii, sector radii, slant length and included angle; it does not create geometry. Pointed cones, partial cone faces and faces with additional boundaries are rejected.

Create an exact area-preserving planar annular-sector Sheet from one complete native conical-frustum face. The profile is placed in the world XY plane with its inner arc starting at originMm; source Solid is preserved. This performs six native history steps (two arcs, two radial lines, join and patch), validates the Sheet and compares exact B-Rep boundary lengths and face area with the source. If interrupted or an error is returned after edits begin, inspect plasticity_status/changes before deciding whether to continue or undo; this tool never retries or rolls back automatically. Pointed cones, partial cone faces, and faces with additional boundary loops are unsupported.

Create independent native Solid or Sheet copies by mapping all faces of selected bodies from one exact source face onto a different exact target face. Source bodies and both reference-face bodies are preserved. Scale U/V/normal and orientation flags are dimensionless native mapping controls; inspect the resulting exact geometry because deformation intentionally changes shape and dimensions.

Create independent native Wire copies by mapping selected curves from one exact source face onto a different exact target face. Source Wires and both reference-face bodies are preserved. Scale U/V/normal and orientation flags are dimensionless native mapping controls; inspect the returned exact curve points, tangents, lengths, and bounds because deformation intentionally changes geometry.

Copy exact current body edges into native Wire curves while preserving their source bodies.

Detach exact current faces from their native shells into separate Sheet bodies.

Insert one separate current fill Sheet into explicit boundary edges of another current target Sheet. Plasticity consumes both input bodies and returns one rebuilt Sheet or Solid in one history step; every prior body and topology reference becomes stale, so inspect the returned state and validate the result.

Explode every face of one or more current multi-face Solid or Sheet bodies into independent single-face native Sheets in one Plasticity history step. The selected bodies are replaced, one result may reuse a source stable ID, and all prior body and topology references become stale; use the returned state.

Create a native Solid from one closed Sheet shell while preserving the source Sheet.

Delete exact current faces from Solid or Sheet bodies, leaving the remaining native shell editable.

Remove selected exact face boundaries by merging the faces into compatible adjacent native surfaces.

Fill one closed boundary loop on a native Sheet using exact current edge IDs.

Cap every planar open boundary of one or more native Sheet bodies in one history step.

Linearly extend selected boundary edges of one native Sheet by a positive millimeter distance.

Create native solid or hollow circular pipes along one or more Wire spines. Diameter and optional wall thickness are millimeters.

Revolve a planar Wire profile around a world-space axis. Axis origin is millimeters and the positive angle is degrees.

Thicken native Sheet bodies into solids. Front and back distances are nonnegative millimeters on opposite sides of each sheet.

Draft exact faces by a signed angle around a planar neutral reference face. Positive angles open away from the reference face.

Offset exact B-Rep faces by a signed distance in millimeters.

Move exact current B-Rep faces by one nonzero world-space delta in millimeters. Plasticity extends and retrims adjacent faces; all topology references become stale after the edit.

Rotate exact current B-Rep faces around a world-space pivot and axis. Pivot is millimeters and angle is degrees; Plasticity extends and retrims adjacent faces, invalidating old topology references.

Scale exact current B-Rep face surfaces by positive world-space XYZ factors about a millimeter pivot. This is a direct surface edit: adjacent faces are retrimmed, and scaling a planar face within its unchanged infinite plane may be a no-op. Re-read exact geometry afterward.

Copy exact current faces from one Solid or Sheet into a new independent native body with nonnegative front and back thicknesses in millimeters. The original body is preserved. Front follows each selected face normal and back goes against it; at least one side must be positive. Re-read all topology and identify the new bodies after success.

Insert native offset loops around exact current faces from one Solid or Sheet at a signed millimeter distance. Positive and negative signs follow Plasticity's face and adjacent-surface orientation; they can place the new loop on the selected face or propagate it over adjacent faces. This splits topology without intentionally changing volume. Re-read the exact new faces and edges instead of assuming a world direction.

Create independent native Sheet patches from exact edge loops selected on one current Solid. The source Solid, including any opening or through-hole, is preserved; this tool constructs covering surfaces and does not claim to heal, fill, or Boolean-close the Solid. Re-read the returned Sheets before thickening, sewing, or other downstream work.

Move exact current B-Rep edges from one body by a nonzero world-space millimeter delta. Plasticity rebuilds adjacent faces, invalidating all prior topology references.

Create native parallel edge offsets on one body at a signed millimeter distance. The sign chooses an adjacent surface according to Plasticity's oriented edge, so re-read the added strip and all topology after the edit.

Remove exact current B-Rep edges from one body through Plasticity's native surface healing. Use for removable split or seam edges between compatible adjacent faces; the native operation can reject structural edges. Re-read all topology after success.

Insert exact native split vertices at one positive millimeter distance along every incident edge of one or more current Solid or Sheet vertices from the same body. This preserves the outer shape and volume while rebuilding topology; it does not move the corner, chamfer it, or fillet it. Re-read every topology reference after success.

Repeat one exact connected feature-face set on the same Solid or Sheet in a native rectangular array. Counts include the source feature and spacing is center-to-center in millimeters. Re-read all topology and validate the result because Plasticity must recognize the selected faces as a repeatable feature.

Repeat one exact connected feature-face set on the same Solid or Sheet around a world-space axis in a native radial array. Count includes the source feature and sweep is in degrees. Re-read all topology and validate the result because Plasticity must recognize the selected faces as a repeatable feature.

Remove selected faces and shell a solid with an inward or outward wall thickness in millimeters.

Turn one or more current Solid bodies into closed hollow Solids without removing an opening face. Inward preserves the outside envelope; outward preserves the original interior envelope. Wall thickness is in millimeters and every body is changed in one native history step.

Mirror exact bodies across a world-space plane, either copying or moving the originals.

Create a native rectangular body pattern using counts and center-to-center spacing in millimeters.

Create a native radial body pattern around a world-space axis.

Distribute current Solid or Sheet bodies along the full length of one native Wire spine. Count includes the source position; Plasticity creates independent native bodies, preserves the spine, and applies its verified tangent-following orientation in one history step.

Rename one body with Undo support.

Delete selected body IDs with Undo support.

Undo the most recent document edit.

Redo the most recent undone document edit.

Set a named viewport orientation and optionally fit all geometry.

Save a PNG screenshot of the visible Plasticity renderer viewport to a new file without overwriting. Refuses hidden windows to avoid stale frames.

Save the current document to a new .plasticity file without overwriting.

Open a .plasticity document after writing the current document to a new backup file.

Export exact B-Rep bodies to a new STEP file without overwriting.

Export exact native B-Rep bodies to a new Parasolid text (.x_t) or binary (.x_b) file without overwriting. Use this for high-fidelity exchange with software that supports the Parasolid kernel format.

Tessellate exact B-Rep bodies to a new millimeter-scaled binary STL for slicing. The result is a derived mesh, while STEP remains the editable source.

Tessellate current exact Solid or Sheet bodies to a new validated 3MF for slicers. Plasticity 26.1.3 declares meters, so the adapter applies its verified 0.001 scale to preserve millimeter dimensions and reports mesh bounds from the saved package. The result is a derived mesh; keep .plasticity or STEP as the editable source.

Tessellate current exact Solid or Sheet bodies to a new validated Wavefront OBJ without overwriting. Coordinates are written in millimeters with Z up; the result reports counts and bounds parsed from the saved file. OBJ is a derived mesh for interchange or rendering, so retain .plasticity, STEP, or Parasolid as the editable source.

Export coplanar native Wire profiles as a new millimeter-scaled SVG without overwriting. B-Rep Lines, full circles, trimmed circular arcs, and native Ellipse segments remain exact when Plasticity exposes their analytic carrier data; non-rational polynomial BCurves of integer degree 1 through 3, including periodic curves, are exported span-by-span as cubic Bezier commands only after additional exact B-Rep validation samples pass. Degree-1 and degree-2 non-rational fixtures have live production stdio verification with independent B-Rep samples on Plasticity 26.1.3. Rational BCurves that fit a conic and agree with 65 dense exact B-Rep samples are represented by an SVG ellipse/arc and marked with sampled-validation metadata; this does not prove global equality, and anything that fails validation uses the approximation path. Other planar B-Rep curves use an adaptive polyline checked at quarter samples against curveChordToleranceMm and curveChordAngleDegrees and are explicitly marked as approximations in SVG metadata. Returned deviation is the maximum tested chord deviation, not a proof of global error. Noncoplanar Wires are rejected; for Solid drawings use plasticity_export_hiddenline_svg. Keep the .plasticity or STEP file as the editable source.

Project current native Solid B-rep geometry through Plasticity's hidden-line generator into a new vector SVG without overwriting. Requires an orthographic current view; output geometry is calibrated to projected model millimeters and includes visible and hidden line styles. Set a named orthographic camera with plasticity_set_view first when needed. This is a derived drawing, not an editable CAD model; retain .plasticity or STEP.

Import exact STEP geometry into the current document. Returns compact document state, imported body IDs/count, source SHA-256, and a persistent local provenance record with native B-Rep measurements. Optionally provide HTTPS source provenance. The record is historical after further edits; use plasticity_body_info for selected bodies and plasticity_measure_solid_properties for measurements.

Download one explicitly selected HTTPS STEP file or ZIP containing exactly one STEP member from a public hostname and import it as exact native CAD geometry. Returns compact document state and changed body IDs/count; use plasticity_body_info for selected exact B-Rep detail. Provide source.sourceUrl as the direct file/archive URL and source.sourcePageUrl as the candidate page URL when available. The downloader pins a resolved public IPv4 address per request, limits redirects, archive size and response size, validates the STEP envelope and archive CRC, stores private content-addressed source and import artifacts, and records both source/archive provenance hashes. Search candidates first with plasticity_search_product_references, review the source, license, format and fit, then select one before calling this importer. Never use arbitrary, untrusted candidates as exact dimensions.

Read one explicitly selected product or CAD source page and list its direct STEP/Parasolid, mesh (including 3MF), and drawing links. This performs one bounded HTTPS GET of the selected HTML page only; it does not download an asset, import geometry, run page scripts, or mutate Plasticity. Provide the same explicit allowedDomains used for the selected search. Page hosts, redirects, and returned asset hosts must match those domains, DNS is pinned to public IPv4, and all query-bearing asset links are omitted to avoid exposing expiring credentials. Review the returned exact asset URL, format, license, and fit before separately choosing a download/import tool.

Download one explicitly selected HTTPS STL or OBJ file from a public hostname and import it into Plasticity as an approximate reference mesh. Declare sourceUnit because STL is unitless and community mesh scale can be ambiguous. The downloader pins a public IPv4 address, limits redirects and payloads to 64 MiB, validates the mesh structure and finite coordinates, and stores a private SHA-256-addressed artifact. The import is journaled with redacted source provenance and measured mesh bounds. This is tessellated reference geometry, not native B-rep or proof of exact product dimensions; do not use it as a fit-critical datum without checking official drawings or user-confirmed measurements.

Download one explicitly selected direct HTTPS 3MF file from a public hostname and import it through Plasticity as approximate reference mesh geometry. The embedded 3MF unit is used by the native importer; returned bounds are reference-mesh measurements, not native B-rep accuracy or proof of exact product dimensions. The downloader pins public IPv4, limits redirects and payloads to 64 MiB, validates the bounded ZIP/XML package, CRCs, paths and mesh indices, then stores a private SHA-256-addressed artifact. It records the artifact hash, source, acquisition and all import-time meshes in the persistent CAD reference-import registry; read the historical record with plasticity_get_cad_reference_import. Review source, license and product/SKU first; this does not import exact editable CAD or start a print.

Import a local SVG as editable native planar Wires in one Plasticity history step. The source unit explicitly defines how SVG coordinate values map to physical length. Closed non-self-intersecting contours also produce revision-bound planar Regions that can be extruded or used as profiles; verify exact Wire geometry after import because SVG transforms and curve content can change the resulting placement and topology.

Import a local STL or OBJ as an approximate reference mesh in one Plasticity history step. The source unit is explicit because STL is unitless and community OBJ scale is often ambiguous. The result can be selected and transformed, but its mesh bounds and triangles are reference evidence only; use manufacturer CAD, drawings, or user-confirmed dimensions for exact modeling.

Import a local 3MF as approximate reference mesh geometry through Plasticity's native importer. The file's embedded unit determines scale; read the returned mesh bounds and keep them distinct from exact native B-rep dimensions. The server stores the artifact hash, import-time document/revision, mesh IDs, bounds and topology counts in the persistent CAD reference-import registry; optional source URLs are redacted and retain license and confidence. Read the historical record with plasticity_get_cad_reference_import.

Import a validated Parasolid text (.x_t) or binary (.x_b) file as native editable B-Rep geometry. Returns compact document state, imported body IDs/count, artifact SHA-256, and historical document/revision provenance with exact imported-body measurements. Optional source metadata records the direct asset and source page. Use plasticity_body_info for selected bodies.

Download one explicitly selected HTTPS Parasolid file or ZIP containing exactly one compatible member from a public hostname and import it as exact native B-Rep geometry. Returns compact document state and changed body IDs/count; use plasticity_body_info for selected exact B-Rep detail. Select the source page and license first; source.sourceUrl is the direct file/archive URL and source.sourcePageUrl should identify the product page. Choose the exact representation explicitly; .xmt_txt is text and .xmt_bin is binary. DNS is pinned to public IPv4 addresses; redirects, archive/response size, content header, revision, hashes and imported B-Rep provenance are checked.

Generate ordered cohesive split planes from the selected exact slicer profile hash, its layer height, the caller-supplied first interlayer plane anchor in current CAD millimeters, and a confirmed global print build direction. The profile hash, layer height, layer count, build direction and anchor are caller-supplied; this helper does not fetch or authenticate a Workbench job, read CAD placement, or prove the anchor lies on the Solid. For a completed Workbench slice, call workbench_slicer_interface_heights with every selected interfaceLayerIndex; copy each returned relativeOffsetMm into interfaceOffsetsMm in the same order. You may also pass the complete selected response metadata and interfaces as depositionPathEvidence; the planner checks job/profile/G-code hashes, selected layers, layer count and relative offsets, then preserves each road-direction summary in the analysis record as provenance. To receive candidate material frames for every deposited layer, call workbench_slicer_layer_path_orientations in batches of up to 32 and combine the selected results, including the final layer, as layerPathEvidence with matching job/profile/source/G-code hashes. For solver-mapped frames, retrieve complete coverage evidence (linear moves and XY G2/G3 circular arcs using I/J offsets or signed R radius; up to 256 layers), provide user-confirmed pathFrameMapping.slicerXDirectionGlobal, and explicitly confirm with roadAxisMapping that the exact-process coupon axis 1 represents the dominant deposited-road direction. P multi-turn arcs, malformed or mixed I/J-and-R arcs, non-XY arc planes, absolute I/J center mode, and G5/G5.1 splines or G5.2/G5.3 NURBS blocks remain partial and cannot qualify this solver mapping. The cohesive analysis additionally requires useOrthotropicBulkProperties and the measured single-material tensor. It then applies that one shared tensor with a per-layer local frame in both Mode-I and Turon; it does not create multiple materials or layer-varying property values. Without roadAxisMapping, returned frames remain candidates and do not affect solver input. The same measured, direction-independent cohesive law is repeated at each interface; individual roads, within-layer raster mixtures and direction-dependent adhesion are not modeled. Interface and per-layer evidence, when both supplied, must refer to the same job and artifacts. These relative offsets are in the slicer's build frame; map only their distances onto the confirmed CAD build axis, and do not assume that absolute slicer coordinates equal CAD coordinates. Before cohesive analysis, record both the immutable Workbench profile hash and its layerHeightMm in the measured interface-test process; the analysis rejects absent or mismatched layer heights. Exact plane intersections are checked only when meshing succeeds. A complete stack up to 255 interfaces is fully analyzed; larger stacks may select up to 255 explicitly chosen interfaces and are marked incomplete. Omitted interfaces are not analyzed and no full-stack delamination conclusion is valid. Pass the returned plan as layerPlanePlan and the returned planes as splitPlanes to plasticity_analyze_cohesive_interface; that tool checks the profile hash and layer height against the measured test record and, for orthotropic bulk, checks the direction against the exact coupon frame.

Run an experimental cohesive solver response for one current native Solid split by 1..255 ordered parallel planes into bulk regions representing the same single printed material. Every new analysis requires a profile-bound layerPlanePlan created by plasticity_plan_cohesive_layer_planes; arbitrary unbound split planes are rejected. The plan profile hash and nominal layer height must match the measured same-material interface-test process, actual G-code relative layer offsets should be copied when available, and orthotropic bulk requires its build axis to match the coupon frame. This tool rejects dissimilar-material bond records before meshing or solving and does not calculate multi-material prints. The default Mode-I route requires a stored same-material layer-failure DCB curve and uses isotropic bulk response with pinned Code_Aster 15.2; modeILaw defaults to CZM_EXP_REG for compatibility, while CZM_LIN_REG must be explicitly selected. This option applies only to Mode-I; mixed-mode requests use the calibrated Turon law. Both Mode-I choices use measured peak traction and integrated fracture energy, and neither reproduces arbitrary measured curve shape. When useOrthotropicBulkProperties is explicitly enabled, Mode-I uses Code_Aster 17.4 and one exact-process coupon's homogeneous orthotropic tensor and confirmed print frame identically on both sides. The mixed-mode Turon route additionally requires same-process ENF and at least two MMB records plus traceable initial cohesive stiffness K in MPa/mm with the exact process identity inside initialStiffnessEvidence.materialProcess, and an explicit displacement vector with both opening and shear components. Its shear component must align within one degree with the shared measured ENF/MMB in-plane axis; unsupported directions are rejected before meshing because CZM_TURON has one tangential law. It may use the same orthotropic single-material tensor. Both require one unambiguous exact-process coupon, one directly evidenced Poisson ratio, and current planar support/load faces. The input carries one material process and one Poisson ratio; solver region labels A and B only identify the two sides of the same material. Multi-plane coverage is complete only when every interface is represented (up to 255); selected planes from taller stacks remain explicitly incomplete. The same measured same-material layer law is repeated at every plane; individual roads and layer-by-layer raster directions are not resolved. Turon interface adhesion is direction-independent in its tangent plane. The tool exports STEP, creates a conforming cohesive mesh with a repeated measured layer law at each requested plane, runs a pinned network-disabled Code_Aster solver, checks the Plasticity revision before and after solving, and saves an immutable report. Code_Aster Mode-I results label V3 semantics explicitly: CZM_EXP_REG reports a damage variable, while CZM_LIN_REG uses V3=2 for a fully broken element; do not read V3 as the same normalized damage fraction for both laws. The response does not establish strength, design adequacy or print approval.

Read a persisted cohesive-interface solver report by ID and report whether its saved CAD session, document, body, revision, immutable interface test, and exact-process coupon records still match current evidence. The saved output remains a raw solver response with mesh-screening diagnostics; it never establishes strength or print approval.

Run bounded linear-elastic CalculiX analyses for one selected native Solid, modeled as one material and one print process; never combine material datasets. The default isotropic model uses youngsModulusMPa and poissonRatio with exact matching evidence. A process-matched uniaxial coupon may supply the isotropic Young's modulus only; use plasticity_match_material_coupon_data first and bind only an unambiguous exact-process record. Poisson ratio always needs separate evidence.

To use an orthotropic model, pass orthotropicMaterial. In that mode youngsModulusMPa is E1 and poissonRatio is nu12; orthotropicMaterial supplies E2, E3, nu13, nu23, G12, G13 and G23. When the immutable exact-process coupon record contains the measured full tensor and nu12, bind it with orthotropicMaterial.couponRecordId and process, and set materialCoupon to the same record ID/process. The server checks E1, nu12, all seven remaining constants, each evidence object and the confirmed print frame against that record before solving and again when the saved report is read. Otherwise, provide separate exact-value evidence for every property. Measured/sourced evidence requires URL, SHA-256 and locator; assumptions require an explicit derivation and remain scenario-only. The server rejects an unstable normal-compliance matrix and non-positive moduli. Give axis1DirectionGlobal, axis2ReferenceDirectionGlobal, and buildDirectionGlobal in the global CAD frame. The build direction must be explicitly user-confirmed or sourced, and material axis 3 must align with it; axis 3 is the layer-normal direction. Axis 1 and 2 are orthogonalized into a right-handed local frame; zero or parallel axes are rejected. Without layerPlanePlan, this is one homogeneous frame. For a per-layer orientation analysis, pass a complete layerPlanePlan made from the exact single-material process and G-code: profile hash and measured layer height must match orthotropicMaterial.process; the plan must include every layer (up to 256), confirmed slicer-to-CAD and coupon-road-axis mappings, and complete planar road-direction evidence for every layer. Linear deposition and XY circular arcs are integrated for I/J offsets and signed R radii. P multi-turn arcs, malformed or mixed arc forms, non-XY arc planes, absolute I/J center mode, and G5/G5.1 splines or G5.2/G5.3 NURBS blocks remain incomplete. The server cuts the STEP B-Rep at the planned interfaces before meshing, verifies each face-to-surface fragment by exact plane/bounds/area evidence, and assigns each conformal volume region the same measured tensor with that layer's mapped frame. The report marks these components as layer-local. Static FEA treats the interfaces as perfectly bonded; it does not predict delamination or interlayer failure. Use the separate cohesive-interface analysis with measured interface tests for that. Varying layer frames cannot use directional component allowables expressed in one fixed material frame. Orthotropic analyses do not accept a von Mises allowable because it is not a qualified orthotropic failure criterion. The screen checks no multiaxial interaction and is diagnostic only; it is not a failure verdict, strengthPass or print approval. Without all nine supported directional limits, no orthotropic stress screen is returned.

For an optional measured 3D Tsai-Wu screen, bind one qualified single-material print record by supplying its ID as orthotropicMaterial.tsaiWuQualificationRecordId together with orthotropicMaterial.process. The server resolves the immutable record only when it is the unique exact match for printer, filament, profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature and measured slicer layer height, and checks that its measured print axes match the FEA material frame. It loads the nine un-factored X/Y/Z tensile/compressive and XY/XZ/YZ shear failure strengths plus three normalized XY/XZ/YZ normal-interaction coefficients and their source evidence from that record; agents do not copy these values manually. Each strength evidence entry must attest testAxis and testMode in the confirmed material frame; each biaxial interaction and its source dependencies must attest its corresponding plane. Legacy records without directional metadata remain readable but cannot be bound to Tsai-Wu FEA. Alternatively, provide the complete inline orthotropicMaterial.tsaiWuCriterion with process-matched, direction-qualified evidence. The interaction matrix must be positive definite. CalculiX evaluates the full material-local tensor at each integration point and reports the maximum failure index and proportional load factor to index one per case and mesh. This diagnostic first-failure surface never establishes whole-part strength or print approval. It assumes one homogeneous material and does not resolve individual roads, discrete layer delamination, nonlinear response, fatigue, buckling or convergence; without measured data it is omitted.

For isotropic analyses only, an optional factoredVonMisesAllowableMPa requires directly measured/sourced matching evidence with URL, SHA-256, locator and factoredVonMisesAllowableBasis. The allowable must already include design factors and apply to this material/process; do not convert a generic datasheet strength or raw coupon peak into an allowable. Its comparison with sampled mesh peaks is diagnostic only and never establishes strength, convergence, strengthPass or print approval.

Choose either legacy supportFaceIds (all three global translations fixed on each planar face) or 1–8 explicit supportConditions; each condition fixes only its listed global translation axes x/y/z to zero at every node on that face. Ask the user to confirm the actual restraints; do not infer them from a photo or face orientation. Before solving, the tool checks that mapped support-node translations remove all six rigid-body translations and rotations; this is only a restraint-rank check and does not prove elastic stability or physical support validity. It has no friction, contact or rotational support model. Provide either legacy single-case loads or up to eight named independent loadCases. Each case may combine uniform face tractions in N/mm² and resultant face loads with force N, application point mm and free moment N·mm. Resultant-load points must lie on their selected planar faces; free moments use a balanced equivalent nodal couple over that face. Optionally request one to three mesh refinement steps, each halving meshSizeMm; the job count is bounded to 12. Cases at the same mesh level are compared only when Gmsh reproduces an identical mesh byte for byte. Each case reports sampled trends for raw maximum von Mises stress and observed displacement (increasing, decreasing, unchanged, non-monotonic, or insufficient-levels), plus raw peak-locator centroid shifts. These are diagnostics only, never a convergence pass, strength pass or print approval.

Read a persisted linear-static FEA report and check whether its CAD binding and referenced exact-process coupon record still match. When an isotropic report contains a directly traceable factored von Mises allowable, returns the raw peak stress comparison for every saved case and mesh level. When a homogeneous orthotropic report contains all nine traceable, already factored directional allowables, returns a componentwise maximum-normal/maximum-shear screen in its material-local frame; the fixed-frame screen is unavailable for layer-local reports. An optional 3D Tsai-Wu screen uses measured single-process strength data and biaxially derived interaction coefficients evaluated at each local integration-point tensor; it reports failure index and proportional load factor only. Layerwise static FEA applies one measured material tensor in G-code-mapped frames and assumes perfectly bonded interfaces; it does not predict delamination. Use cohesive-interface analysis with measured interface tests for that. These results remain diagnostic and never establish strength, convergence or print approval. Missing, corrupted or newly conflicting coupon records make a bound report stale; reports remain immutable and are not refreshed automatically.

Compare 2–4 saved static FEA reports from the same CAD revision, body, supports, loads, material evidence and native geometry. It merges only byte-identical repeated mesh hashes with matching solver results, then reports sampled stress/displacement trends across the combined levels. If all inputs share a directly traceable factored von Mises allowable, returns its raw peak screen; if they share all nine orthotropic factored directional allowables, returns the material-local componentwise maximum-stress screen. The orthotropic screen assumes one homogeneous continuum; it does not model layer interfaces, delamination or different-material joints. Both omit unresolved failure modes and are diagnostic only, not convergence estimates, strength verdicts or print approval. CAD or material evidence freshness is reported separately.

Calculate nominal peak stress for each directly loaded physical specimen as measured peak force in N divided by measured net cross-section in mm² (N/mm² = MPa). Preserves per-specimen failure location and hashed source locator; only specimens confirmed to fail at the printed interface contribute to the descriptive range, mean and sample standard deviation. Other failures remain individually visible and are excluded; no interface failures yields null summary values. These values are not local interface tractions, design allowables, statistically qualified bounds or a cohesive law.

Calculate an exploratory Mode-I G_I-versus-crack-length curve from caller-selected DCB crack-growth observations using Modified Beam Theory (MBT). Provide the exact single-material print process, global layer-interface normal, test method, test protocol SHA-256/date, each specimen's measured width/length/arm thickness, at least three strictly increasing observed crack lengths, corresponding positive force and machine-compliance-corrected load-point displacement, source SHA-256 and per-point source locator. Explicitly attest quasi-static linear-elastic behavior. Rows with displacement/crack-length above 0.4 are rejected because large-displacement correction is not implemented. This is not a standards-conformance determination; ASTM D5528 states a scope for unidirectional fiber-reinforced polymer composites. It does not calculate a traction-separation curve, cohesive law, design allowable, or Creality material property; only specimens with caller-confirmed interface failure are marked eligible for same-material interlayer fracture evidence. The tool is read-only and does not persist tests.

Calculate an exploratory ENF Mode-II initiation energy from caller-supplied compliance-calibration results. For each specimen provide at least three distinct crack lengths and compliances taken from the inverse initial-linear force-displacement slope, using the same specimen support/loading fixture as the fracture run; provide the measured initial crack and peak force, exact one-material process, global interface normal and perpendicular global ENF shear direction, protocol hash/date, source SHA-256 and locator for each calibration and fracture input, plus explicit linear-elastic/quasi-static and calibration attestations. The tool fits C = A + ma^3 and evaluates G_IIc = 3mPc^2a0^2/(2*b), retaining fit R-squared and each source. It does not interpret raw machine traces, correct compliance, determine ASTM validity, or claim ASTM D7905 conformity (that standard's scope is unidirectional carbon/glass fiber-reinforced laminates; printed PLA is outside the validated scope). This is a read-only exploratory energy estimate, not an R-curve, traction-separation law, cohesive parameter, design allowable or Creality material property; only caller-confirmed layer-interface failures are eligible for same-material interlayer fracture evidence.

Calculate an exploratory mixed-mode initiation-energy partition from manually measured MMB critical force, crack length, specimen/fixture geometry, exact one-material process, interface normal and interface-plane shear direction. Requires source hashes/locators for each specimen and measured flexural modulus plus orthotropic E11/E22/G13 evidence with an explicitly confirmed mapping (axis 1 = shear, axis 2 = in-plane transverse, axis 3 = interface normal). Requires the caller to confirm lever self-weight is measured negligible or counterbalanced. Uses the Reeder-Crews beam-theory equations to return Mode-I, Mode-II and total energy-release rates and the Mode-II energy fraction. It does not select initiation from raw test traces, establish ASTM D6671 conformity (printed PLA is outside that laminate standard's validated scope), infer a traction-separation law, qualify material, or authorize design/printing. Only confirmed interface failures are eligible as same-material interface-energy evidence; this read-only estimate is separate from the measured-curve registry and cohesive FEA.

Read-only preview of caller-selected physical MMB initiation forces in one explicitly mapped local UTF-8 CSV. Reads only a regular non-symlink file up to 16 MiB and 250,000 data records. For each specimen, map its ID and force columns/units/sign, then select the exact CSV record associated with the manually determined crack-initiation criterion and provide measured crack length, geometry, failure location, exact one-material process, sourced same-process flexural/orthotropic moduli, and confirmed material-axis mapping. The importer converts force units only, preserves the CSV SHA-256 and exact record locator, and returns the MMB energy-partition preview. It never selects a peak or identifies initiation from raw curves, and does not validate the MMB fixture or ASTM D6671 conformity. The preview does not register physical evidence, produce a traction-separation curve/cohesive law, qualify material, or authorize design/printing.

Persist an immutable caller-confirmed physical MMB initiation-energy partition for one exact single-material process, interface normal, in-plane shear axis, protocol and date. Requires traceable source evidence for the manually selected initiation force, specimen geometry, same-process flexural/orthotropic moduli and confirmed material axes, plus explicit observed failure location and initiation criterion. The server recomputes the exploratory Reeder-Crews estimate. This separate MMB energy registry is not a traction-separation registry or cohesive-FEA input; a calculated energy partition is not a cohesive law, peak traction, or design allowable.

Find caller-confirmed physical MMB initiation-energy records only for an exact single-material process, interface normal, in-plane shear direction and protocol SHA-256. Returns no-match, matched or ambiguous evidence. Matching exploratory MMB energy is not a traction-separation curve, cohesive-law calibration or design allowable.

List immutable caller-confirmed physical MMB initiation-energy records as compact summaries. Use plasticity_read_mmb_mode_i_ii_energy_test for all measured inputs, initiation criteria, source hashes/locators, moduli evidence and server-recomputed results. These are not cohesive laws and are not applied to FEA.

Read one immutable physical MMB energy record including measured geometry, initiation criterion/force, source evidence, material-axis mapping, moduli, and the server-recomputed exploratory Mode-I/Mode-II partition. It is not a traction-separation law or cohesive-FEA input.

Read-only preview of raw ENF Mode-II force/displacement data in one explicitly mapped local CSV. For each specimen, manually group at least three distinct compliance-calibration runs by specimen/run ID and crack length, then select the exact record numbers in each initial-linear force-displacement region. Also select the physical initiation/peak record from its fracture run; the tool never chooses a peak or finds a linear region for you. It fits displacement versus force for each selected calibration run, converts the resulting compliance and selected fracture force into the caller-specified units, preserves per-source SHA-256/record locators, and calculates an exploratory Mode-II initiation-energy preview. Explicitly map columns, units, signs, CSV formatting, same-material process, interface normal and in-plane shear direction; caller attestations do not independently verify machine-compliance handling, fixture identity, or failure location. It does not register a physical test or claim ASTM D7905 conformity, an R-curve, cohesive law or design allowable. Review every selected run and fit before any physical-evidence recording.

Persist an immutable caller-confirmed physical ENF Mode-II initiation-energy test for one exact single-material process, interface normal, in-plane shear direction, protocol and date. Requires compliance calibration from the same fixture and source SHA-256/locator for every calibration and fracture input; recomputes the exploratory C-a^3 energy result server-side. This independent Mode-II energy registry is not a traction-separation registry and is not used by cohesive FEA; do not treat G_IIc as a peak shear traction, cohesive stiffness, R-curve or design allowable.

Find recorded physical ENF Mode-II initiation-energy tests only for an exact single-material print process, interface normal, in-plane shear direction and test-protocol SHA-256. Returns no-match, matched or ambiguous evidence. This exploratory energy match is not a traction-separation curve or cohesive FEA calibration.

List immutable caller-confirmed physical ENF Mode-II energy records as compact summaries. Use plasticity_read_enf_mode_ii_energy_test for all calibration/fracture measurements, source hashes/locators and fit diagnostics. These are not cohesive laws and are not applied to FEA.

Read one immutable physical ENF Mode-II energy record including all compliance calibration runs, fracture inputs, source provenance, fit diagnostics, and the server-recomputed initiation-energy result. It is not a traction-separation law or cohesive FEA input.

Read-only preview of an explicitly mapped local DCB Mode-I raw force/displacement CSV. For every physically observed crack-growth point, the caller must select the exact CSV record number and supply its observed crack length; map the specimen, force and displacement columns, units, signs, delimiter and decimal separator explicitly. Converts only force/displacement units, retains the source SHA-256 and record locators, and calculates the same exploratory MBT G_I-versus-crack-length preview. It does not filter acquisition samples, infer crack growth, choose peaks, correct machine compliance, register a physical test or establish ASTM conformity, a cohesive law or a design allowable. Review the preview against the physical log, then call plasticity_record_dcb_mode_i_energy_test only after the physical test is confirmed.

Persist an immutable physical DCB Mode-I energy-release test for one exact single-material process. Requires caller confirmation that all inputs are physical, the observed crack ran on the same-material layer interface, displacement is machine-compliance-corrected load-point displacement, and the test was quasi-static/linear-elastic. Recomputes the exploratory MBT curve server-side; records the protocol hash, date, global interface normal, per-specimen failure location and source evidence. This separate energy registry is not a traction-separation registry and is not read by cohesive FEA; do not use it as peak traction, cohesive stiffness, or a design allowable.

Find recorded physical DCB Mode-I energy tests only for an exact single-material printer/profile/process, global interface normal and test-protocol SHA-256. Returns no-match, matched or ambiguous evidence; a match is not a cohesive FEA calibration or design allowable.

List immutable physical DCB Mode-I energy records as compact summaries. Use plasticity_read_dcb_mode_i_energy_test to retrieve full measured observations and MBT results for a selected record. These records are separate from traction-separation data and are not applied to cohesive FEA.

Read one content-addressed physical DCB Mode-I energy record, including the retained source locators and derived MBT G_I versus crack-length results. This is not a traction-separation law or cohesive FEA input.

Read a caller-selected local UTF-8 CSV containing direct tensile-coupon force samples, find the maximum sampled tensile force for each explicitly listed specimen, and return a preview with source SHA-256 plus nominal force/area stress. The caller must map exact CSV headers, delimiter, decimal separator, force unit and force sign, and supply each specimen's measured net cross-section and observed failure location; none are inferred. Reads only a regular non-symlink file up to 16 MiB, leaves it unchanged, and does not register a physical test. This is a peak-strength screen only: it does not filter or compliance-correct machine data, create DCB/ENF/MMB traction-separation curves, estimate fracture energy or cohesive parameters, or establish an allowable. Review the preview and then explicitly call plasticity_record_material_interface_test to persist caller-attested physical evidence.

Read-only preview of per-specimen DCB Mode-I, ENF Mode-II or MMB mixed-mode traction-separation curves from a caller-selected local CSV. The caller must explicitly map specimen and measurement columns, units, decimal/delimiter settings, and attest that the input already contains physical compliance-corrected separations and tractions; raw force-displacement data are rejected. The importer checks complete measured curves, summarizes their peak and integrated work, and preserves the CSV SHA-256 plus exact record locators. It reads only a regular non-symlink UTF-8 file up to 16 MiB, does not alter the file, register a test, infer failure location, correct compliance, calculate a cohesive law, or establish a design allowable. Review specimen, fixture, process, failed interface and correction method before explicitly recording each accepted specimen with plasticity_record_material_interface_test.

Store an immutable caller-attested physical test of a printed layer interface for one material and one exact print process shared by both sides. Supply one materialProcess with printer/material/profile/orientation/infill percentage and pattern/wall-loop/top-bottom-shell/nozzle-temperature/measured-layer-height identity, the global interface normal, test mode, load direction and a hashed specimen/fixture protocol. For a direct peak-strength series, attach one specimenResults entry per physical sample with measured peak force, net cross-section, individual failure location and traceable source; the selected representativeSpecimenId must match measuredPeakStrengthMPa and its evidence. Use plasticity_calculate_interface_specimen_strengths to derive nominal N/mm² = MPa values and descriptive sample statistics. A direct-strength record without a traction-separation curve must include these raw specimen results. For a curve-based fracture test, explicitly set fractureMethod to dcb-mode-i, enf-mode-ii or mmb-mixed-mode; loading direction and a free-text testMethod do not establish the fracture method. DCB requires normal-tension loading and a scalar curve, ENF requires interface-shear loading and a scalar curve, and MMB requires mixed-mode loading and a vector curve. When available, attach depositionPathEvidence from the actual sliced specimen G-code with matching profile, source/G-code hashes, per-layer road direction summaries and user-confirmed slicer-to-global axes; this is provenance only and is not converted into adhesion strength. Normal-tension loads must align with the interface normal; interface-shear loads must lie in its plane; mixed-mode loads must contain both components. Curves require source SHA-256 and locator, begin at zero, end at zero traction, and match the evidenced measured peak. This stores test evidence only: it does not derive design allowables or approve a design. Repeating identical data is idempotent.

Find physical layer-interface tests only for one exact printer/material/profile/orientation/infill percentage and pattern/wall-loop/top-bottom-shell/nozzle-temperature/measured-layer-height process, test mode, interface normal, load direction and hashed test protocol. Returns no-match, matched or ambiguous. Legacy records without the full process structure remain readable but cannot satisfy a new exact match. A match is experimental evidence only; it does not become a design allowable or a structural-analysis pass.

List immutable caller-attested same-material printed layer-interface test records with one exact print-process identity, setup, failure location and traceable measurement evidence, including full scalar pure-mode or vector mixed-mode traction-separation curves. Results are not automatically applied to static FEA or design checks.

Summarize a stored, complete measured traction-separation curve only when its physical method is explicitly classified as DCB Mode I, ENF Mode II or MMB, consistently with its load mode. Pure-mode records report peak, first-segment stiffness and fracture energy; mixed-mode records report normal/tangential work, mode-mix energy fraction and resultant stiffness. Legacy curves without an explicit fracture method remain readable but are not analyzed. These are measured-curve summaries only, not qualified cohesive-law parameters or design allowables.

Fit a candidate Code_Aster CZM_TURON ETA_BK only from immutable measured Mode-I DCB, Mode-II ENF and at least two distinct-ratio MMB test records for one same-material printed-layer interface and its exact single print process; different-material bond tests are rejected. Each testMethod must explicitly identify DCB, ENF or MMB. Rejects missing curves, non-interface failure, conflicting exact-setup records or mismatched processes/interface normals. Returns measured pure-mode peak tractions and per-MMB energy residuals for engineering review. It does not identify the initial stiffness K, qualify a cohesive law, authorize FEA, establish a design allowable or approve a part.

Plan physical measurements for one selected single-material print process and solver scope; no multi-material calculation is supported. Returns required directional coupon, biaxial or same-material layer-interface evidence without inventing property values, material properties or allowables. For DCB it includes a clearly labeled generic-PLA literature geometry/acquisition precedent, not a Creality property, normative specimen size, or sample-count requirement; applicability and specimen sizing must be checked for the exact process and fixture. A focused layer-interface-normal-tension scope plans only a direct peak-strength test; it does not replace Mode-I DCB fracture evidence or calibrate a cohesive law. A focused Mode-II scope plans an ENF compliance-calibration initiation-energy estimate only; it does not claim ASTM D7905 conformity for printed PLA or provide a cohesive input curve. Coupon records require measured E1; ask for other properties only when the selected solver scope needs them. Layer-interface tests concern cohesion between layers of that same material. Initial cohesive stiffness K is MPa/mm evidence supplied directly to plasticity_analyze_cohesive_interface, not stored in the MPa interface-peak registry. Requires exact printer/material/profile/orientation/infill-percentage-and-pattern/wall-loops/top-and-bottom-shell-layers/nozzle-temperature/measured-layer-height identity; print-axis mapping and interface directions must be confirmed.

Store immutable physical coupon data for one exact single-material printer/material/profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature, and measured layer height. E1 (youngModulusMPa) with measured evidence is required. Isotropic shear modulus and tensile/shear strengths are optional and must each be paired with evidence; record them only when measured for a selected analysis. Optionally include measured Poisson ratio nu12 with its evidence IDs. Add E2/E3, nu13/nu23, G12/G13/G23 under orthotropicMaterial; its propertyEvidence IDs point to exact measured entries in evidence[]. Include the three global print axes; axis 3 must align with the build direction, and orientation.evidence must be user-confirmed or traceable. An optional tsaiWuCriterion can store nine directly measured directional failure strengths and three derived normalized interaction coefficients from biaxial tests. Each test evidence entry must state testAxis and testMode in the confirmed material frame (X/Y/Z tension or compression, XY/XZ/YZ shear, and corresponding-plane biaxial interactions); interaction evidence dependencies must carry the same biaxial plane. These attestations are stored with each exact evidence entry and the process in the immutable record hash. This model uses one material per part; it does not model multi-material prints. These un-factored strengths are not design allowables or proof of part strength.

Explicitly consolidate 2–8 compatible immutable physical coupon records for one exact single-material print process. It combines only non-conflicting measured properties and their source evidence; it never averages or infers values. Supply specimenCount as the caller-confirmed number of unique physical specimens across all source records. The result is a new immutable record with composedFromRecordIds, and can be used for exact-process matching and FEA binding. Different processes, conflicting values or print frames are rejected.

Match coupon data only for an exact printer, material, profile SHA-256, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature, and layer height. Returns no-match, matched or ambiguous; conflicting isotropic properties, orthotropic tensors, Tsai-Wu strengths/interactions or print axes are never selected silently. A single record may be selected over compatible records that contain only a subset of its measured values. Complementary partial records remain ambiguous; call plasticity_combine_material_coupon_data with their IDs only after the user confirms the unique physical specimen count. The combine tool never averages or infers values and rejects conflicts. A match returns its measured/derived evidence and dependencies for review. Coupon strengths are not design allowables.

List immutable caller-attested material coupon records with their source evidence, exact process identity and any composedFromRecordIds. This physical-test registry is separate from slicer material names, densities and temperatures.

Store immutable caller-attested physical test results for a multi-hole printed plate with the exact printer/material/profile/orientation, measured specimen dimensions and hole layout, load axis, fastener clearance, clamp condition and fixture configuration. Every peak load must link to measured N evidence and the measured geometry/test report must be traceable. This records observations only: it does not derive design allowables, infer a shared-ligament formula, qualify statistical strength or approve a part. Repeating identical data is idempotent.

Find physical multi-hole joint tests only for an exact print process, rectangular specimen dimensions, hole centers/diameters, fixture configuration, load axis, fastener diameter/clearance and clamp condition. Returns no-match, matched or ambiguous records; reordering hole input does not change the match. A match is test evidence only, not a capacity, design allowable, strength pass or proof that a different part/support setup is equivalent.

List immutable caller-attested physical multi-hole joint test records. The registry preserves measured failure loads and observed modes with exact process/geometry/fixture and report evidence; it does not calculate design allowables or certify a part.

List deterministic member, plate and section methods plus Codex analysis availability. Search accessible primary product/material sources before asking the user for known facts. Before requesting missing print-material data, call plasticity_plan_single_material_strength_tests for the selected method and exact one-material process. Use its measurement matrix to ask only for evidence needed by that route; explain why the measurements matter and never invent coupon values or allowables. A DCB task may contain an explicitly labeled generic-PLA literature geometry as a starting reference only; do not present it as a Creality property, normative specimen size, or sample-count requirement, and require checking the selected process, fixture and method. Keep the confirmed road/build axes and same-material layer-interface assumptions explicit. For Creality PLA literature references, read plasticity://strength/interlayer-literature-baseline. It contains separate CR-PLA and Hyper PLA records plus generic-PLA Z-tension/DCB context; do not merge product families or transfer values across SKUs. It is a screening reference only: do not register literature values as physical coupons or interface tests, bind them to the user's exact K1C process, treat them as design allowables, or infer a traction-separation curve from fracture energy. Manufacturer-mirror discrepancies and non-matching printer/process tests must remain visible; ask for exact-process physical tests before a calibrated cohesive result. When the user provides raw direct-tension machine CSV, call plasticity_import_interface_tensile_csv first with explicit specimen-ID/force headers, delimiter, decimal separator, force unit and sign, measured net area and observed failure location for every coupon. Review its hashed peak-force preview; the importer does not filter or correct machine data, register a test, or derive DCB/ENF/MMB curves. For directly loaded specimens supplied in another format, record every sample's measured peak force, net cross-section, observed failure location and SHA-256/source locator, then call plasticity_calculate_interface_specimen_strengths; report force/area as nominal specimen stress. Include only confirmed interface failures in its summary statistics; show bulk/fixture failures individually and do not pool them with interface failures. Keep all statistics descriptive only. Do not call these local interface tractions or design allowables, and do not use them as a cohesive law. When a raw DCB force/displacement CSV is available, use plasticity_import_dcb_mode_i_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting explicitly, then select the exact source record for every manually observed crack-growth point and enter that measured crack length; never let the importer filter traces, infer growth or select peak loads. It converts units only, preserves the source hash/record locators and calculates an exploratory MBT G_I(a) curve. Require the caller to establish machine-compliance-corrected load-point displacement and quasi-static linear-elastic behavior; attestations are not independent verification. Review selected rows, crack lengths, calculation and limitations against the physical log. If no CSV is available, plasticity_calculate_dcb_mode_i_energy accepts equivalent manually selected observations with traceable source hash/locators. This does not conform the test to ASTM D5528 (whose stated scope is unidirectional fiber composites), create a traction-separation law, or qualify CR-PLA. Never convert G_I into peak traction or strength without separate evidence. When raw ENF force/displacement CSV is available, use plasticity_import_enf_mode_ii_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting; group calibration rows into each measured run/crack length, manually select only the initial-linear records, and manually select the observed initiation/peak record for the fracture run. The tool fits compliance from the selected rows, then the existing ENF calculator fits C against a^3 and returns exploratory G_IIc; it never searches for the linear range, crack growth or peak. Review each compliance fit, source hash/record locator, force/displacement correction and fixture log. If pre-reduced compliances are supplied, use plasticity_calculate_enf_mode_ii_energy directly. Neither path determines ASTM D7905 validity for printed PLA, registers physical evidence, builds an R-curve or yields a cohesive law. Keep the confirmed in-plane shear direction exact; do not merge ENF runs with differing direction just because their layer normals match. When the requested failure mode is layer separation, distinguish CR-PLA's physical tensile/infill study and the two-sample CR-PLA-associated vertical layer-adhesion screen from the Hyper PLA flat tensile/flexural study: observed flexural delamination is qualitative failure-mode evidence, not a measured interface law. The baseline also records an upright generic-PLA tensile coupon on a Creality Ender 3 Pro and a custom generic-PLA interface coupon/calibrated cohesive parameter; neither identifies the filament as Creality nor establishes a same-process cohesive calibration. For only a nominal normal-strength screen across the layers, use the focused layer-interface-normal-tension scope; require measured failure at the interface and do not use that peak as a cohesive law. Use layer-interface-mode-i for a Mode-I fracture response, layer-interface-mode-ii for exploratory ENF Mode-II initiation energy using a same-fixture compliance calibration, and layer-interface-mixed-mode only when the requested analysis needs DCB/ENF/MMB cohesive evidence. The focused ENF calculator fits the experimentally supplied compliance against crack length cubed and uses the initiation peak; it is not an R-curve, does not verify the physical fixture or raw compliance fit, and does not claim ASTM D7905 validity for printed PLA. Layerwise static FEA assumes perfectly bonded interfaces and cannot answer a delamination question. After reviewing an exploratory DCB energy preview against the physical test log, persist it only through plasticity_record_dcb_mode_i_energy_test with explicit confirmation that the measurements came from real physical tests. Retrieve a full record with plasticity_read_dcb_mode_i_energy_test and require exact process/interface-normal/protocol matching through plasticity_match_dcb_mode_i_energy_test before comparing runs. For ENF Mode-II energy, record reviewed measurements only through plasticity_record_enf_mode_ii_energy_test with the same physical-test confirmation; retrieve all calibration/fracture inputs with plasticity_read_enf_mode_ii_energy_test and exact-match process/interface-normal/in-plane-shear-axis/protocol using plasticity_match_enf_mode_ii_energy_test. Both immutable energy registries are separate from peak-strength and traction-separation records and are not consumed by cohesive FEA. For an exploratory mixed-mode initiation partition, use plasticity_calculate_mmb_mode_i_ii_energy only when measured MMB force and geometry, same-process flexural and orthotropic moduli, and material-axis mapping are available; require lever weight to be measured negligible or counterbalanced. This Reeder-Crews beam-theory estimate does not establish ASTM D6671 validity for printed PLA and does not replace full mixed-mode traction-separation curves or provide a cohesive law. For raw MMB force CSV, use plasticity_import_mmb_mode_i_ii_energy_csv only with manually selected initiation records and a physically observed criterion; it does not search traces for onset/peak. Record a preview only after reviewing the source and confirming the data are from physical tests; matching caller confirmation with plasticity_record_mmb_mode_i_ii_energy_test persists and server-recomputes this separate energy estimate. Use plasticity_match_mmb_mode_i_ii_energy_test only for the exact process, interface normal, shear axis and protocol; read the full evidence with plasticity_read_mmb_mode_i_ii_energy_test. This registry is not cohesive input or an FEA source. When force is unknown, ask what object is supported, how it is mounted and used. Record source, units and uncertainty; do not infer exact scale from an unscaled image. Ask at most one next-step question package per response, then wait for the user's answer. Include only facts needed to choose the next safe step; defer material, manufacturing, tolerances and detailed dimensions until they affect that decision. Re-evaluate after every answer and ask a focused follow-up only when it changes the method, required evidence or next action. If the user does not know, move to one useful contextual clue such as the supported object, use, environment or mounting; do not repeat a list of unknowns or guess. On a new bracket task, first ask compactly what it supports, its load/use and how it is mounted; defer section, material/process and displacement questions until that first answer narrows the load path and method. Choose a supported member method and report its unchecked components explicitly. For a flat rectangular panel, establish net pressure and the real condition of all four edges before choosing the simply-supported plate method. Do not infer edge support from appearance. For a straight prismatic rectangular member in centred axial compression, use the Euler column method only when effective-length factor K, elastic limit, compressive allowable, and the actual restraint condition are supported by evidence. Use the weakest section axis, require a negative force value for compression, and reject Euler results when its predicted critical stress exceeds the supplied elastic limit; do not guess K or treat the method as an inelastic, eccentric-load, local-buckling, or whole-part check. For an integral rectangular enclosure wall, inspect two opposed planar faces on the same Solid with plasticity_inspect_integral_rectangular_plate, then use plasticity_verify_integral_plate_strength to replace dimensions from exact native B-rep. The inspector accepts only rectangular faces whose outlines match or inset by one wall thickness on each side. It measures geometry only: establish the real support, pressure, material and edge conditions separately; never assume a box wall is simply supported. For section analysis, identify the critical plane and explain the load path that makes it critical. Use an existing planar face when it matches; otherwise inspect an arbitrary plane through the Solid. For one fastener carrying in-plane plate load, establish the load direction and separate bearing, shear and tensile allowables before calculating. Never substitute compressive strength for bearing strength. For a fastener group on a rectangular planar face, inspect exact hole centers and boundaries, then check center-to-edge distance, hole-edge clearance, pitch, ligament and every applicable head, washer, nut or driver envelope against explicitly sourced or user-approved criteria. A measured layout without criteria is not a pass, and a layout pass is not a strength pass. When checking the fastener member, establish its grade, tensile stress area, effective shear area at the actual plane, one or two shear planes, total axial tension including applicable preload, and actual shear load. Explain why the selected combined-load interaction is applicable before accepting it. For a tapped hole, nut, or threaded insert under axial load, establish the designation, pitch, actual engagement, fully formed engaged thread count, and configuration-matched allowable loads for internal-thread stripping, external-thread stripping, and fastener tension. Do not infer any capacity from nominal M size. A procured nut or insert requires a specified assembly allowable or dedicated test rather than an unqualified nominal shear-area calculation. For solver-backed static FEA, first use plasticity_match_material_coupon_data when a physically tested material record may match the print process. Only an unambiguous exact match for printer, material, profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature and measured slicer layer height can supply Young's modulus; pass the selected record ID/process and the exact recorded modulus so the server binds and validates it. If match is ambiguous because compatible partial records have no single consolidated record, ask the user for the confirmed count of unique physical specimens and call plasticity_combine_material_coupon_data with those exact record IDs; it preserves measured values and evidence without averaging. If records conflict, stop and ask the user which physical measurement applies. For an orthotropic print model, do not use a uniaxial coupon as a full tensor. When an immutable exact-process record contains the complete measured tensor and nu12, pass orthotropicMaterial.couponRecordId and process, plus materialCoupon with the same record ID; the MCP checks E1, nu12, every remaining constant, each evidence object and the confirmed global print axes, and rechecks the record when reading the saved report. Otherwise pass E2/E3, nu13/nu23, G12/G13/G23 with separate exact evidence, global material axes 1 and 2, and a user-confirmed or sourced global build direction; material axis 3 must align with that build direction and represents the homogeneous layer-normal response. This does not represent individual layers or prove adhesion. Model one material and one print process per analyzed part; do not combine material datasets or infer a multi-material print. When the selected exact-process coupon record contains a qualified Tsai-Wu dataset, pass its ID as orthotropicMaterial.tsaiWuQualificationRecordId along with the identical orthotropicMaterial.process; the server loads measured strengths, biaxial interaction data and source evidence, then verifies the axes. Do not copy or reconstruct criterion values by hand. If the exact-process match remains ambiguous after consolidation, ask the user to resolve conflicting measurements; never choose a record silently. The legacy youngsModulusMPa and poissonRatio fields are E1 and nu12. Ask the user to resolve the print-axis orientation if it is unknown; never infer it from a photo or CAD face. The server checks positive-definite compliance and rejects a von Mises allowable or coupon binding for this orthotropic model. Otherwise attach youngsModulusEvidence whose value exactly matches the modulus; measured/sourced evidence needs a URL, SHA-256 and locator. poissonRatioEvidence is always required and must exactly match the ratio. A physical coupon record may store measured nu12 with its exact process. If using its Tsai-Wu record-binding path, the server verifies the nu12 value and evidence against that record; otherwise supply the input evidence directly from a traceable source. Never infer nu12 from a generic plastic label. An assumed value must include its reason and stay explicitly scenario-only after discussing the assumption with the user. Optionally supply a directly measured or sourced, process-applicable factored von Mises design allowable using factoredVonMisesAllowableMPa, exact matching factoredVonMisesAllowableEvidence with URL, SHA-256 and locator, and factoredVonMisesAllowableBasis. It must already include the design factors. Never turn a generic tensile strength or raw coupon peak into a design allowable. The report will compare every sampled raw mesh peak with it as a diagnostic screen only: above means a sampled peak exceeds that supplied allowable; below does not prove strength. This never changes strengthPass or print approval. Confirm the actual restraints with the user. Legacy supportFaceIds fix all three global translations on each selected planar face; use explicit supportConditions only when the user has specified which global x/y/z translations are zero on each face. Each condition acts on all nodes of that face, is not a frictionless-contact or rotational support, and may leave rigid-body modes or create a singular system; do not infer it from a photo or face normal. Supports and loaded faces must not share mesh nodes. Every selected planar load face needs an explicit uniform traction and/or resultant force with its point of application and free moment. Use plasticity_analyze_static_fem only for one Solid with 1–8 explicit support conditions and load vectors grounded in a selected planar face. Use named loadCases for physically distinct scenarios such as weight, operating force and handling load; do not combine mutually exclusive scenarios. Each case has a separate solver result, and comparison proceeds only when the generated meshes are byte-identical. Set meshRefinementSteps to 1–3 for two to four mesh levels when a trend across successively halved element sizes is useful; the report classifies the sampled direction of raw maximum von Mises stress and observed displacement as increasing, decreasing, unchanged, non-monotonic or insufficient-levels. For more than four levels, run separate analyses against the same current CAD revision and call plasticity_compare_static_fem_refinement_reports; it merges only reports with matching loads, supports, material evidence, native geometry and byte-identical overlapping mesh results. Check the returned freshness before relying on it. Treat all trend labels and relative changes as diagnostic evidence only. Never call a mesh trend a pass or proof of convergence. Each result includes the raw maximum C3D4 integration-point stress element and its mesh-element centroid in millimetres; this is a mesh-bound locator, not an averaged stress field, a resolved critical-region boundary, or proof of a physical hotspot. Locations may move between refinement levels. The legacy top-level load fields still represent one case. The tool returns total and per-support reaction forces and moments, global force/moment equilibrium residuals and a named displacement axis for each case; per-support values are diagnostic resultants over each support node set. Re-read it with plasticity_static_fem_report; any CAD revision change makes it stale.

For layerwise static FEA, keep the single-material exact process consistent from measurement through solver input. Read the immutable profile hash, infill pattern, wall loops, top/bottom shell layers and measured layer height from workbench_manufacturing_profiles, then slice the intended model/orientation with the selected Creality K1C profile. Record actual specimen settings if per-object overrides differ from profile defaults; an unchanged profile hash does not erase those differences. Call workbench_slicer_layer_path_orientations in batches of up to 32 and combine layer-path results for every deposited layer, including the final deposited layer, into layerPathEvidence with identical job ID, profile hash, source-artifact hash and G-code hash. Layerwise static orientation is supported only for a complete linear or planar circular-arc direction result for every layer in stacks of 2–33 layers; do not fill missing or curved-path directions by inference. Confirm pathFrameMapping.slicerXDirectionGlobal in CAD global coordinates and explicitly confirm that the same exact-process coupon's material axis 1 represents the dominant deposited-road direction in roadAxisMapping. Call plasticity_plan_cohesive_layer_planes with the same profile hash and layer height, current CAD anchor at the first interlayer plane, confirmed CAD build direction, total layer count, and the full layer-path evidence/mappings; pass its returned plan as layerPlanePlan to plasticity_analyze_static_fem. When the slicer supplies actual interface heights, call workbench_slicer_interface_heights for every interface in this complete stack, set each interfaceOffsetsMm value to its depositionLayerZMm minus firstDepositionLayerZMm, and attach the matching job/profile/source/G-code hashes as depositionPathEvidence; this preserves first-layer and adaptive heights instead of assuming nominal uniform spacing. The static analyzer requires the plan to match the one measured orthotropic process and applies the same measured orthotropic tensor to each layer with its confirmed G-code-mapped frame. It assumes perfectly bonded layers and cannot assess delamination. Do not use static FEA to assess delamination; use the separate same-material cohesive route only when matching physical interface tests are available. Never infer layer directions, material identity or interlayer strength from a photo, generic material label or nominal slicer preset.

For orthotropic FEA, optionally supply all nine directly traceable, already factored X/Y/Z tensile and compressive plus XY/XZ/YZ shear limits in orthotropicMaterial.factoredAllowables, with separate exact-value evidence and an applicability/design-factor basis. Never substitute generic datasheet strength or an unqualified coupon peak. The returned componentwise maximum-stress screen uses local material-axis stress extrema and is diagnostic only; it assumes one homogeneous orthotropic continuum, does not model layer interfaces, delamination or different-material joints, and omits multiaxial interaction. Matched interlayer-test data can inform Z-tension and XZ/YZ-shear allowables but does not turn this into a cohesive-interface analysis. Never report this screen as verified layer adhesion, part strength or print approval. For an optional 3D Tsai-Wu first-failure screen in orthotropic linear FEA, require one exact single-material printer/material/profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers and nozzle-temperature identity in orthotropicMaterial.process. Prefer binding the unique exact-process qualification by its orthotropicMaterial.tsaiWuQualificationRecordId; the server loads its nine directly measured, un-factored X/Y/Z tensile/compressive and XY/XZ/YZ shear failure strengths, three normalized XY/XZ/YZ normal-interaction coefficients and evidence, then verifies material axes. Each strength test must attest its material-frame axis and mode (for example, x tension is material-1 tension; xy shear is material-1-2 shear); each interaction and its source dependencies must attest the corresponding biaxial plane. Legacy records without these direction attestations remain readable but cannot qualify a Tsai-Wu FEA. A complete inline orthotropicMaterial.tsaiWuCriterion remains supported when needed. Interactions must be derived from traceable biaxial tests. Ask the user for these records if missing; never copy generic datasheet values, assume the conventional interaction coefficient or infer it from uniaxial coupons. The normalized interaction matrix must be positive definite. The result reports local integration-point failure indices and proportional load factors to index one only. A load factor is not a design safety factor, and neither a sub-unity index nor a large load factor means the part passed. The model still represents one homogeneous material, not individual roads or delamination.

When actual test-coupon G-code is available, preserve its profile/source/G-code hashes, selected per-layer road summaries and explicitly user-confirmed slicer-to-global axes in depositionPathEvidence. This is test provenance only, not an adhesion measurement or solver input; never reconstruct road direction from nominal slicer settings.

When physical adhesion between printed layers of one material is relevant, use plasticity_record_material_interface_test only for caller-provided measured test results with the same exact printer/material/profile process on both sides, interface normal, test mode, load direction, fixture/specimen protocol hash and observed failure location. Normal-tension load direction must align with the interface normal; interface-shear load direction must lie in its plane; mixed-mode must contain both components. Store full compliance-corrected pure-mode curves as scalar separation/traction data and MMB curves as separate normal/tangential separation and traction components, with source SHA-256 and locator. Call plasticity_analyze_material_interface_test_curve to summarize measured work and mode mixity. When the user has a CSV of already processed physical fracture data, use plasticity_import_interface_fracture_csv for a read-only per-specimen preview; explicitly map the method, columns, units and CSV formatting, and attest that the values are already compliance-corrected physical traction-separation data. Never convert raw machine force-displacement data with this importer. Verify each source hash/record locator, specimen, fixture, exact print process and observed failure plane before separately recording any curve with plasticity_record_material_interface_test. For a CZM_TURON candidate fit, provide Mode-I DCB, Mode-II ENF and at least two MMB tests at distinct measured energy fractions; all ENF and MMB tests must use the same in-plane shear axis within one degree because this route has a single tangential cohesive law. The MCP requires those protocol identifiers in each testMethod and reports the pure-mode peaks plus fit residuals. Review residuals against test uncertainty. Do not invent ETA_BK. K is not identified by that fit and must not be silently guessed. Code_Aster CZM_TURON uses one normal and one tangential cohesive response, sharing tangential strength and fracture energy across both in-plane tangent directions; it cannot represent direction-dependent shear adhesion. Matching shear axes prevents mixing directional datasets but does not prove that the bond is isotropic. These outputs summarize physical evidence only; they are not qualified cohesive-law parameters, design allowables or a part FEA. Do not claim bond integrity from the homogeneous orthotropic FEA screen. Use plasticity_analyze_cohesive_interface only for a single-material printed part, with an exact-process coupon for that one bulk material, traceable Poisson ratio and a matching immutable same-material layer test. Dissimilar-material printed bonds are rejected before meshing and are outside this calculation scope. Mode-I analysis requires a full normal-tension DCB traction-separation curve with failure at the layer interface. Its modeILaw may explicitly select CZM_EXP_REG or CZM_LIN_REG; omitted input preserves the CZM_EXP_REG default. Both parameterize their softening law from measured peak traction and integrated fracture energy and do not fit the full measured curve shape. Review this choice against the measured curve and record the reason; do not infer it from part geometry. Read the returned solver.result.v3Interpretation: for CZM_EXP_REG, V3 is a damage variable in [0,1]; for CZM_LIN_REG, V3=2 means the cohesive element is completely broken, so do not present it as the same normalized damage fraction. By default it uses isotropic bulk elasticity with pinned Code_Aster 15.2. If useOrthotropicBulkProperties is explicitly enabled, Mode-I uses Code_Aster 17.4 and applies one measured homogeneous orthotropic tensor; when explicit layerwise mapping is enabled, each bulk layer uses its G-code-mapped frame while all layers share that same tensor. The mixed-mode Turon route additionally requires same-process-pair DCB, ENF and at least two MMB tests at distinct measured energy fractions, traceable K, and an explicit global displacement with both opening-normal and in-plane tangential components greater than one degree. The displacement's shear axis must match the common measured ENF/MMB shear axis within one degree; reject other directions because this solver law has one tangential response. All ENF and MMB tests must use the same in-plane shear axis within one degree. Supply 1..32 ordered parallel split planes for a same-material layer stack; their normals may be tilted in global coordinates, and the same measured layer law is repeated at every plane. This equivalent-interface assumption does not resolve individual roads or within-layer raster variation. Either route may accept useOrthotropicBulkProperties when the exact-process coupon contains measured E2/E3, nu13/nu23, G12/G13/G23 evidence and a confirmed or traceable global print frame. Check the exact-process coupon match is unambiguous and do not infer axes from a photo or CAD face. Confirm the tested material's side relative to the chosen split-plane normal; do not infer this from the body or face normals. The measured test direction must match the normal/shear mode; the support face lies below the first split plane and the loaded face above the last plane along the shared normal. The server derives peak traction, integrated fracture energy and displacement endpoint from exact tests, exports the selected current Solid, creates a conforming multi-region mesh with a cohesive element set at every requested plane, and aborts if the CAD revision changes. Review the mesh-resolution screen and perform refinement/sensitivity work as needed. Turon approximates measured curves using peak/area parameters and its K still needs sensitivity analysis. The bulk tensor remains one measured single-material continuum whose local frame may vary by mapped layer; repeated cohesive planes use the same measured, direction-independent interface law and do not resolve individual deposited roads, within-layer raster mixtures or direction-dependent adhesion. Use results only as raw solver responses; never report it as a part-strength verdict, qualified layer-adhesion value, print approval or design allowable. Before asking for a physical coupon record, obtain the selected immutable profile hash plus material.nozzleTemperatureC, slicer.layerHeightMm, slicer.nominalInfillPercent, slicer.sparseInfillPattern, slicer.wallLoops, slicer.topShellLayers, and slicer.bottomShellLayers from workbench_manufacturing_profiles when those settings are available. Carry the exact process values into the record; do not infer missing infill or substitute a generic profile. These are resolved profile defaults; sparse fill or shell settings may be overridden per object and must not be assumed when a modifier was used. If the coupon was printed with a different setting, first register/select the matching immutable process profile and hash. When layer positions should follow the actual print, record the measured profile hash and layer height in the physical interface-test and material-coupon process; layer height is part of exact process identity. After slicing, if the job reports a complete deposition-height schedule, call workbench_slicer_interface_heights for every selected interface index. Derive each offset as that interface’s depositionLayerZMm minus firstDepositionLayerZMm and pass the ordered values as interfaceOffsetsMm; this preserves first-layer and adaptive heights. Also pass the selected interface response plus its job/profile/source/G-code hashes, layer count and coordinate frame as depositionPathEvidence so the cohesive report retains the exact per-layer road-orientation observations. This evidence preserves the measured toolpath but does not qualify material properties. Use workbench_slicer_layer_path_orientations in batches of up to 32 to retrieve every layer direction (up to 33 total layers) when the user wants layerwise solver orientation. Confirm how slicer X maps into the CAD global frame and that the exact-process coupon axis 1 represents the dominant deposited-road direction; pass those confirmations as pathFrameMapping and roadAxisMapping. Combine complete responses, including the final deposited layer, as layerPathEvidence with shared job/profile/source/G-code hashes. Every layer must have complete linear or planar circular-arc coverage; do not treat unsupported arc/spline moves as complete. With explicit roadAxisMapping and useOrthotropicBulkProperties, Mode-I and Turon use one measured tensor with a separate local frame per layer. Without roadAxisMapping, frames remain candidates and do not affect solver response. This does not model multiple materials, layer-varying properties, within-layer raster mixtures or directional interface adhesion. Returned coordinates are in slicer build coordinates: map only relative offsets onto the confirmed CAD print axis, never copy absolute slicer Z into CAD. Otherwise the planner uses the nominal profile height. Call plasticity_plan_cohesive_layer_planes with that same profile hash and layer height, a point on the first interlayer plane after object placement, confirmed global build direction, total layer count and explicit interface indices. Copy both its plan and planes into layerPlanePlan and splitPlanes; analysis rejects legacy test records without a recorded layer height and any height that differs from the measured process profile. It also checks profile identity, plane coordinates, and (for orthotropic bulk) alignment with the coupon’s confirmed build direction. For up to 32 interfaces the planner requires the complete stack; for larger stacks it marks selected planes as incomplete, and omitted interfaces remain unanalyzed. Do not present selected-plane analysis as full-stack delamination resistance. If no validated placement/anchor is available, ask instead of inferring layer positions from an image or display mesh. Read maximumVonMisesElementSICN alongside each raw stress-peak locator to see the Gmsh SICN of that exact tetrahedron. Compare it with the mesh minimum only as local mesh-shape context; neither a high value nor separation from the minimum proves stress accuracy, a resolved hotspot, convergence, or strength. Read maximumPrincipalStressMPa and minimumPrincipalStressMPa as raw tensile/compressive principal extrema across sampled integration points, each with its own mesh locator. Their refinement trends and signed relative changes expose mesh sensitivity only; they are diagnostic and must never be compared with a von Mises allowable or reported as a pass without a criterion qualified for the selected failure mode. The public FEA tool measures the rank of all fixed global translations at the actual mapped mesh nodes and stops before CalculiX if they leave any of the six rigid-body translation/rotation modes unconstrained. A full rank of six is necessary to remove those rigid-body modes, but it does not establish physical support validity, elastic stability, or absence of local mechanisms. For a heat-set insert, use pullout and torque-out capacity only when the evidence matches the exact insert, host material, print profile, orientation, pocket and installation process. Ask for the worst-case demand on one insert; do not divide a group load evenly without a load-path model. For two or more fasteners under an in-plane load, establish every transfer-point coordinate, both force components, the point of application and any free moment. Use the elastic group method only after confirming a rigid attachment and identical in-plane fastener stiffness. Use each fastener’s own vector resultant for any member check. On a rectangular mounting face, call plasticity_check_fastener_group_layout with an explicit opposedFaceId to verify matching native perforated faces and exact plate thickness. This is geometry evidence only and does not establish a load path or capacity. plasticity_verify_fastener_group_plate_bearing compares each elastic per-fastener demand with a directly traceable, configuration-matched, already factored bearing allowable using exact measured thickness and hole diameter. It can additionally check a straight transverse net-tension section only when you provide the external tensile resultant separately, identify local X or Y as its axis, supply a distinct traceable factored tensile allowable, and confirm uniform membrane tension, centered through-thickness loading and a straight transverse failure path. Never substitute per-fastener demands for the external plate tension or infer that resultant from a sketch. Optionally use edgeShearOut with a separate traceable, factored shear allowable to check local two-plane tear-out for per-fastener vectors aligned to local X or Y; diagonal vectors and e/d below 1.5 are unsupported, while e/d below 2 remains conditional. Angled/staggered fracture paths, compression-side buckling, shared-ligament interaction, unsupported tear-out directions, bypass and complete-joint strength remain unchecked; even a within-allowable result is only a conditional local screen. If a physical multi-hole plate test is run, record the measured specimen, exact hole layout, print-process/profile, fixture, load axis, individual peak loads and observed failure modes with plasticity_record_fastener_group_test, then use plasticity_match_fastener_group_test to find only an exact configuration match. A test match is evidence only: it does not produce a design allowable or pass a different part or support setup. To add a test benchmark to the CAD-bound report, first confirm that the record's exact process and fixture/load path apply to the current part. Supply the selected immutable record ID, an independently evidenced dimensional equivalence tolerance, the exact process, safety factor, and explicit process/fixture confirmations; the server then checks every measured plate and hole dimension against the live B-rep. This comparison only checks factored external tensile demand against the lowest observed specimen peak. It is not a statistically reduced allowable, strength pass/fail, or proof for unobserved failure modes. Do not select a nearby test by appearance or silently assume fixture equivalence. The single-through-fastener plate method assumes one hole; never repeat it per hole and aggregate the results into a multi-hole plate or whole-joint pass. Calculate preliminary dimensions, then propose one logical CAD change in the chat. Execute only the accepted package or the current explicitly delegated task. Delegation ends when the task ends and is not restored after restart. Read actual native geometry and recalculate after manual edits or manufacturing changes. Workbench is optional. Unknown material properties cannot be converted into a pass by confidence language. For an end-to-end tongue-root scenario, use plasticity_calculate_tongue_root_strength_from_coupon_data or plasticity_verify_tongue_root_strength_from_coupon_data. They match the physical coupon record to the exact Workbench profile hash, printer, material, orientation, infill, nozzle temperature and measured slicer layer height, then copy only measured Young's/shear moduli and their evidence into the calculation. Supply separately sourced tensile/shear design allowables and explain their applicability; raw coupon strengths are never substituted for allowables. These tools keep material suitability unconfirmed and the result conditional. No-match or conflicting records return without a calculation. If no record exists, ask for the report and unresolved process details, and record it only after the user confirms the physical tests. The registry does not check test-standard compliance, derive statistical design values, or certify a part. Never substitute a generic filament label or slicer properties for coupon evidence. For axial rectangular or rectangular-cantilever scenarios with exact-process coupon data, use plasticity_calculate_rectangular_strength_from_coupon_data. It copies only measured Young's modulus and its evidence. Provide independent tensile design allowable evidence, plus an independent compressive allowable for a cantilever; state their applicability basis. Never map raw coupon tensile/compressive strengths into design allowables. These scenarios remain conditional with material suitability unconfirmed. Exact-process no-match or ambiguity returns without saving a calculation. Nominal section stress is not whole-part validation.

Report whether isolated Codex live web search for product CAD and dimensional references is available. This capability has no CAD, shell, local-file, or MCP access.

Search live web sources for candidate CAD models and reliable dimensioned references using an isolated Codex profile. Results are unverified discovery leads only: source-page and direct asset URLs are cross-checked against Codex web-search/open-page results, but licensing, paid/account access, fit, source quality and dimensions still require review. Each candidate reports accessStatus separately from licenseStatus; a paid product page is not a direct asset URL and must not be passed to an importer. It never downloads/imports files or mutates CAD. Prefer manufacturer sources; specify allowedDomains when discovery should be scoped. searchTimeoutMs defaults to 180000 and accepts 30000–180000 ms for slow searches.

Run one isolated Codex turn for factual extraction and the single next decision-relevant question package. Include each prior full question with its user answer on follow-up turns. This writes a durable request record, contacts Codex, never mutates CAD and never authorizes a print. Search accessible primary product/material sources before asking the user for known facts. Before requesting missing print-material data, call plasticity_plan_single_material_strength_tests for the selected method and exact one-material process. Use its measurement matrix to ask only for evidence needed by that route; explain why the measurements matter and never invent coupon values or allowables. A DCB task may contain an explicitly labeled generic-PLA literature geometry as a starting reference only; do not present it as a Creality property, normative specimen size, or sample-count requirement, and require checking the selected process, fixture and method. Keep the confirmed road/build axes and same-material layer-interface assumptions explicit. For Creality PLA literature references, read plasticity://strength/interlayer-literature-baseline. It contains separate CR-PLA and Hyper PLA records plus generic-PLA Z-tension/DCB context; do not merge product families or transfer values across SKUs. It is a screening reference only: do not register literature values as physical coupons or interface tests, bind them to the user's exact K1C process, treat them as design allowables, or infer a traction-separation curve from fracture energy. Manufacturer-mirror discrepancies and non-matching printer/process tests must remain visible; ask for exact-process physical tests before a calibrated cohesive result. When the user provides raw direct-tension machine CSV, call plasticity_import_interface_tensile_csv first with explicit specimen-ID/force headers, delimiter, decimal separator, force unit and sign, measured net area and observed failure location for every coupon. Review its hashed peak-force preview; the importer does not filter or correct machine data, register a test, or derive DCB/ENF/MMB curves. For directly loaded specimens supplied in another format, record every sample's measured peak force, net cross-section, observed failure location and SHA-256/source locator, then call plasticity_calculate_interface_specimen_strengths; report force/area as nominal specimen stress. Include only confirmed interface failures in its summary statistics; show bulk/fixture failures individually and do not pool them with interface failures. Keep all statistics descriptive only. Do not call these local interface tractions or design allowables, and do not use them as a cohesive law. When a raw DCB force/displacement CSV is available, use plasticity_import_dcb_mode_i_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting explicitly, then select the exact source record for every manually observed crack-growth point and enter that measured crack length; never let the importer filter traces, infer growth or select peak loads. It converts units only, preserves the source hash/record locators and calculates an exploratory MBT G_I(a) curve. Require the caller to establish machine-compliance-corrected load-point displacement and quasi-static linear-elastic behavior; attestations are not independent verification. Review selected rows, crack lengths, calculation and limitations against the physical log. If no CSV is available, plasticity_calculate_dcb_mode_i_energy accepts equivalent manually selected observations with traceable source hash/locators. This does not conform the test to ASTM D5528 (whose stated scope is unidirectional fiber composites), create a traction-separation law, or qualify CR-PLA. Never convert G_I into peak traction or strength without separate evidence. When raw ENF force/displacement CSV is available, use plasticity_import_enf_mode_ii_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting; group calibration rows into each measured run/crack length, manually select only the initial-linear records, and manually select the observed initiation/peak record for the fracture run. The tool fits compliance from the selected rows, then the existing ENF calculator fits C against a^3 and returns exploratory G_IIc; it never searches for the linear range, crack growth or peak. Review each compliance fit, source hash/record locator, force/displacement correction and fixture log. If pre-reduced compliances are supplied, use plasticity_calculate_enf_mode_ii_energy directly. Neither path determines ASTM D7905 validity for printed PLA, registers physical evidence, builds an R-curve or yields a cohesive law. Keep the confirmed in-plane shear direction exact; do not merge ENF runs with differing direction just because their layer normals match. When the requested failure mode is layer separation, distinguish CR-PLA's physical tensile/infill study and the two-sample CR-PLA-associated vertical layer-adhesion screen from the Hyper PLA flat tensile/flexural study: observed flexural delamination is qualitative failure-mode evidence, not a measured interface law. The baseline also records an upright generic-PLA tensile coupon on a Creality Ender 3 Pro and a custom generic-PLA interface coupon/calibrated cohesive parameter; neither identifies the filament as Creality nor establishes a same-process cohesive calibration. For only a nominal normal-strength screen across the layers, use the focused layer-interface-normal-tension scope; require measured failure at the interface and do not use that peak as a cohesive law. Use layer-interface-mode-i for a Mode-I fracture response, layer-interface-mode-ii for exploratory ENF Mode-II initiation energy using a same-fixture compliance calibration, and layer-interface-mixed-mode only when the requested analysis needs DCB/ENF/MMB cohesive evidence. The focused ENF calculator fits the experimentally supplied compliance against crack length cubed and uses the initiation peak; it is not an R-curve, does not verify the physical fixture or raw compliance fit, and does not claim ASTM D7905 validity for printed PLA. Layerwise static FEA assumes perfectly bonded interfaces and cannot answer a delamination question. After reviewing an exploratory DCB energy preview against the physical test log, persist it only through plasticity_record_dcb_mode_i_energy_test with explicit confirmation that the measurements came from real physical tests. Retrieve a full record with plasticity_read_dcb_mode_i_energy_test and require exact process/interface-normal/protocol matching through plasticity_match_dcb_mode_i_energy_test before comparing runs. For ENF Mode-II energy, record reviewed measurements only through plasticity_record_enf_mode_ii_energy_test with the same physical-test confirmation; retrieve all calibration/fracture inputs with plasticity_read_enf_mode_ii_energy_test and exact-match process/interface-normal/in-plane-shear-axis/protocol using plasticity_match_enf_mode_ii_energy_test. Both immutable energy registries are separate from peak-strength and traction-separation records and are not consumed by cohesive FEA. For an exploratory mixed-mode initiation partition, use plasticity_calculate_mmb_mode_i_ii_energy only when measured MMB force and geometry, same-process flexural and orthotropic moduli, and material-axis mapping are available; require lever weight to be measured negligible or counterbalanced. This Reeder-Crews beam-theory estimate does not establish ASTM D6671 validity for printed PLA and does not replace full mixed-mode traction-separation curves or provide a cohesive law. For raw MMB force CSV, use plasticity_import_mmb_mode_i_ii_energy_csv only with manually selected initiation records and a physically observed criterion; it does not search traces for onset/peak. Record a preview only after reviewing the source and confirming the data are from physical tests; matching caller confirmation with plasticity_record_mmb_mode_i_ii_energy_test persists and server-recomputes this separate energy estimate. Use plasticity_match_mmb_mode_i_ii_energy_test only for the exact process, interface normal, shear axis and protocol; read the full evidence with plasticity_read_mmb_mode_i_ii_energy_test. This registry is not cohesive input or an FEA source. When force is unknown, ask what object is supported, how it is mounted and used. Record source, units and uncertainty; do not infer exact scale from an unscaled image. Ask at most one next-step question package per response, then wait for the user's answer. Include only facts needed to choose the next safe step; defer material, manufacturing, tolerances and detailed dimensions until they affect that decision. Re-evaluate after every answer and ask a focused follow-up only when it changes the method, required evidence or next action. If the user does not know, move to one useful contextual clue such as the supported object, use, environment or mounting; do not repeat a list of unknowns or guess. On a new bracket task, first ask compactly what it supports, its load/use and how it is mounted; defer section, material/process and displacement questions until that first answer narrows the load path and method. Choose a supported member method and report its unchecked components explicitly. For a flat rectangular panel, establish net pressure and the real condition of all four edges before choosing the simply-supported plate method. Do not infer edge support from appearance. For a straight prismatic rectangular member in centred axial compression, use the Euler column method only when effective-length factor K, elastic limit, compressive allowable, and the actual restraint condition are supported by evidence. Use the weakest section axis, require a negative force value for compression, and reject Euler results when its predicted critical stress exceeds the supplied elastic limit; do not guess K or treat the method as an inelastic, eccentric-load, local-buckling, or whole-part check. For an integral rectangular enclosure wall, inspect two opposed planar faces on the same Solid with plasticity_inspect_integral_rectangular_plate, then use plasticity_verify_integral_plate_strength to replace dimensions from exact native B-rep. The inspector accepts only rectangular faces whose outlines match or inset by one wall thickness on each side. It measures geometry only: establish the real support, pressure, material and edge conditions separately; never assume a box wall is simply supported. For section analysis, identify the critical plane and explain the load path that makes it critical. Use an existing planar face when it matches; otherwise inspect an arbitrary plane through the Solid. For one fastener carrying in-plane plate load, establish the load direction and separate bearing, shear and tensile allowables before calculating. Never substitute compressive strength for bearing strength. For a fastener group on a rectangular planar face, inspect exact hole centers and boundaries, then check center-to-edge distance, hole-edge clearance, pitch, ligament and every applicable head, washer, nut or driver envelope against explicitly sourced or user-approved criteria. A measured layout without criteria is not a pass, and a layout pass is not a strength pass. When checking the fastener member, establish its grade, tensile stress area, effective shear area at the actual plane, one or two shear planes, total axial tension including applicable preload, and actual shear load. Explain why the selected combined-load interaction is applicable before accepting it. For a tapped hole, nut, or threaded insert under axial load, establish the designation, pitch, actual engagement, fully formed engaged thread count, and configuration-matched allowable loads for internal-thread stripping, external-thread stripping, and fastener tension. Do not infer any capacity from nominal M size. A procured nut or insert requires a specified assembly allowable or dedicated test rather than an unqualified nominal shear-area calculation. For solver-backed static FEA, first use plasticity_match_material_coupon_data when a physically tested material record may match the print process. Only an unambiguous exact match for printer, material, profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature and measured slicer layer height can supply Young's modulus; pass the selected record ID/process and the exact recorded modulus so the server binds and validates it. If match is ambiguous because compatible partial records have no single consolidated record, ask the user for the confirmed count of unique physical specimens and call plasticity_combine_material_coupon_data with those exact record IDs; it preserves measured values and evidence without averaging. If records conflict, stop and ask the user which physical measurement applies. For an orthotropic print model, do not use a uniaxial coupon as a full tensor. When an immutable exact-process record contains the complete measured tensor and nu12, pass orthotropicMaterial.couponRecordId and process, plus materialCoupon with the same record ID; the MCP checks E1, nu12, every remaining constant, each evidence object and the confirmed global print axes, and rechecks the record when reading the saved report. Otherwise pass E2/E3, nu13/nu23, G12/G13/G23 with separate exact evidence, global material axes 1 and 2, and a user-confirmed or sourced global build direction; material axis 3 must align with that build direction and represents the homogeneous layer-normal response. This does not represent individual layers or prove adhesion. Model one material and one print process per analyzed part; do not combine material datasets or infer a multi-material print. When the selected exact-process coupon record contains a qualified Tsai-Wu dataset, pass its ID as orthotropicMaterial.tsaiWuQualificationRecordId along with the identical orthotropicMaterial.process; the server loads measured strengths, biaxial interaction data and source evidence, then verifies the axes. Do not copy or reconstruct criterion values by hand. If the exact-process match remains ambiguous after consolidation, ask the user to resolve conflicting measurements; never choose a record silently. The legacy youngsModulusMPa and poissonRatio fields are E1 and nu12. Ask the user to resolve the print-axis orientation if it is unknown; never infer it from a photo or CAD face. The server checks positive-definite compliance and rejects a von Mises allowable or coupon binding for this orthotropic model. Otherwise attach youngsModulusEvidence whose value exactly matches the modulus; measured/sourced evidence needs a URL, SHA-256 and locator. poissonRatioEvidence is always required and must exactly match the ratio. A physical coupon record may store measured nu12 with its exact process. If using its Tsai-Wu record-binding path, the server verifies the nu12 value and evidence against that record; otherwise supply the input evidence directly from a traceable source. Never infer nu12 from a generic plastic label. An assumed value must include its reason and stay explicitly scenario-only after discussing the assumption with the user. Optionally supply a directly measured or sourced, process-applicable factored von Mises design allowable using factoredVonMisesAllowableMPa, exact matching factoredVonMisesAllowableEvidence with URL, SHA-256 and locator, and factoredVonMisesAllowableBasis. It must already include the design factors. Never turn a generic tensile strength or raw coupon peak into a design allowable. The report will compare every sampled raw mesh peak with it as a diagnostic screen only: above means a sampled peak exceeds that supplied allowable; below does not prove strength. This never changes strengthPass or print approval. Confirm the actual restraints with the user. Legacy supportFaceIds fix all three global translations on each selected planar face; use explicit supportConditions only when the user has specified which global x/y/z translations are zero on each face. Each condition acts on all nodes of that face, is not a frictionless-contact or rotational support, and may leave rigid-body modes or create a singular system; do not infer it from a photo or face normal. Supports and loaded faces must not share mesh nodes. Every selected planar load face needs an explicit uniform traction and/or resultant force with its point of application and free moment. Use plasticity_analyze_static_fem only for one Solid with 1–8 explicit support conditions and load vectors grounded in a selected planar face. Use named loadCases for physically distinct scenarios such as weight, operating force and handling load; do not combine mutually exclusive scenarios. Each case has a separate solver result, and comparison proceeds only when the generated meshes are byte-identical. Set meshRefinementSteps to 1–3 for two to four mesh levels when a trend across successively halved element sizes is useful; the report classifies the sampled direction of raw maximum von Mises stress and observed displacement as increasing, decreasing, unchanged, non-monotonic or insufficient-levels. For more than four levels, run separate analyses against the same current CAD revision and call plasticity_compare_static_fem_refinement_reports; it merges only reports with matching loads, supports, material evidence, native geometry and byte-identical overlapping mesh results. Check the returned freshness before relying on it. Treat all trend labels and relative changes as diagnostic evidence only. Never call a mesh trend a pass or proof of convergence. Each result includes the raw maximum C3D4 integration-point stress element and its mesh-element centroid in millimetres; this is a mesh-bound locator, not an averaged stress field, a resolved critical-region boundary, or proof of a physical hotspot. Locations may move between refinement levels. The legacy top-level load fields still represent one case. The tool returns total and per-support reaction forces and moments, global force/moment equilibrium residuals and a named displacement axis for each case; per-support values are diagnostic resultants over each support node set. Re-read it with plasticity_static_fem_report; any CAD revision change makes it stale.

For layerwise static FEA, keep the single-material exact process consistent from measurement through solver input. Read the immutable profile hash, infill pattern, wall loops, top/bottom shell layers and measured layer height from workbench_manufacturing_profiles, then slice the intended model/orientation with the selected Creality K1C profile. Record actual specimen settings if per-object overrides differ from profile defaults; an unchanged profile hash does not erase those differences. Call workbench_slicer_layer_path_orientations in batches of up to 32 and combine layer-path results for every deposited layer, including the final deposited layer, into layerPathEvidence with identical job ID, profile hash, source-artifact hash and G-code hash. Layerwise static orientation is supported only for a complete linear or planar circular-arc direction result for every layer in stacks of 2–33 layers; do not fill missing or curved-path directions by inference. Confirm pathFrameMapping.slicerXDirectionGlobal in CAD global coordinates and explicitly confirm that the same exact-process coupon's material axis 1 represents the dominant deposited-road direction in roadAxisMapping. Call plasticity_plan_cohesive_layer_planes with the same profile hash and layer height, current CAD anchor at the first interlayer plane, confirmed CAD build direction, total layer count, and the full layer-path evidence/mappings; pass its returned plan as layerPlanePlan to plasticity_analyze_static_fem. When the slicer supplies actual interface heights, call workbench_slicer_interface_heights for every interface in this complete stack, set each interfaceOffsetsMm value to its depositionLayerZMm minus firstDepositionLayerZMm, and attach the matching job/profile/source/G-code hashes as depositionPathEvidence; this preserves first-layer and adaptive heights instead of assuming nominal uniform spacing. The static analyzer requires the plan to match the one measured orthotropic process and applies the same measured orthotropic tensor to each layer with its confirmed G-code-mapped frame. It assumes perfectly bonded layers and cannot assess delamination. Do not use static FEA to assess delamination; use the separate same-material cohesive route only when matching physical interface tests are available. Never infer layer directions, material identity or interlayer strength from a photo, generic material label or nominal slicer preset.

For orthotropic FEA, optionally supply all nine directly traceable, already factored X/Y/Z tensile and compressive plus XY/XZ/YZ shear limits in orthotropicMaterial.factoredAllowables, with separate exact-value evidence and an applicability/design-factor basis. Never substitute generic datasheet strength or an unqualified coupon peak. The returned componentwise maximum-stress screen uses local material-axis stress extrema and is diagnostic only; it assumes one homogeneous orthotropic continuum, does not model layer interfaces, delamination or different-material joints, and omits multiaxial interaction. Matched interlayer-test data can inform Z-tension and XZ/YZ-shear allowables but does not turn this into a cohesive-interface analysis. Never report this screen as verified layer adhesion, part strength or print approval. For an optional 3D Tsai-Wu first-failure screen in orthotropic linear FEA, require one exact single-material printer/material/profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers and nozzle-temperature identity in orthotropicMaterial.process. Prefer binding the unique exact-process qualification by its orthotropicMaterial.tsaiWuQualificationRecordId; the server loads its nine directly measured, un-factored X/Y/Z tensile/compressive and XY/XZ/YZ shear failure strengths, three normalized XY/XZ/YZ normal-interaction coefficients and evidence, then verifies material axes. Each strength test must attest its material-frame axis and mode (for example, x tension is material-1 tension; xy shear is material-1-2 shear); each interaction and its source dependencies must attest the corresponding biaxial plane. Legacy records without these direction attestations remain readable but cannot qualify a Tsai-Wu FEA. A complete inline orthotropicMaterial.tsaiWuCriterion remains supported when needed. Interactions must be derived from traceable biaxial tests. Ask the user for these records if missing; never copy generic datasheet values, assume the conventional interaction coefficient or infer it from uniaxial coupons. The normalized interaction matrix must be positive definite. The result reports local integration-point failure indices and proportional load factors to index one only. A load factor is not a design safety factor, and neither a sub-unity index nor a large load factor means the part passed. The model still represents one homogeneous material, not individual roads or delamination.

When actual test-coupon G-code is available, preserve its profile/source/G-code hashes, selected per-layer road summaries and explicitly user-confirmed slicer-to-global axes in depositionPathEvidence. This is test provenance only, not an adhesion measurement or solver input; never reconstruct road direction from nominal slicer settings.

When physical adhesion between printed layers of one material is relevant, use plasticity_record_material_interface_test only for caller-provided measured test results with the same exact printer/material/profile process on both sides, interface normal, test mode, load direction, fixture/specimen protocol hash and observed failure location. Normal-tension load direction must align with the interface normal; interface-shear load direction must lie in its plane; mixed-mode must contain both components. Store full compliance-corrected pure-mode curves as scalar separation/traction data and MMB curves as separate normal/tangential separation and traction components, with source SHA-256 and locator. Call plasticity_analyze_material_interface_test_curve to summarize measured work and mode mixity. When the user has a CSV of already processed physical fracture data, use plasticity_import_interface_fracture_csv for a read-only per-specimen preview; explicitly map the method, columns, units and CSV formatting, and attest that the values are already compliance-corrected physical traction-separation data. Never convert raw machine force-displacement data with this importer. Verify each source hash/record locator, specimen, fixture, exact print process and observed failure plane before separately recording any curve with plasticity_record_material_interface_test. For a CZM_TURON candidate fit, provide Mode-I DCB, Mode-II ENF and at least two MMB tests at distinct measured energy fractions; all ENF and MMB tests must use the same in-plane shear axis within one degree because this route has a single tangential cohesive law. The MCP requires those protocol identifiers in each testMethod and reports the pure-mode peaks plus fit residuals. Review residuals against test uncertainty. Do not invent ETA_BK. K is not identified by that fit and must not be silently guessed. Code_Aster CZM_TURON uses one normal and one tangential cohesive response, sharing tangential strength and fracture energy across both in-plane tangent directions; it cannot represent direction-dependent shear adhesion. Matching shear axes prevents mixing directional datasets but does not prove that the bond is isotropic. These outputs summarize physical evidence only; they are not qualified cohesive-law parameters, design allowables or a part FEA. Do not claim bond integrity from the homogeneous orthotropic FEA screen. Use plasticity_analyze_cohesive_interface only for a single-material printed part, with an exact-process coupon for that one bulk material, traceable Poisson ratio and a matching immutable same-material layer test. Dissimilar-material printed bonds are rejected before meshing and are outside this calculation scope. Mode-I analysis requires a full normal-tension DCB traction-separation curve with failure at the layer interface. Its modeILaw may explicitly select CZM_EXP_REG or CZM_LIN_REG; omitted input preserves the CZM_EXP_REG default. Both parameterize their softening law from measured peak traction and integrated fracture energy and do not fit the full measured curve shape. Review this choice against the measured curve and record the reason; do not infer it from part geometry. Read the returned solver.result.v3Interpretation: for CZM_EXP_REG, V3 is a damage variable in [0,1]; for CZM_LIN_REG, V3=2 means the cohesive element is completely broken, so do not present it as the same normalized damage fraction. By default it uses isotropic bulk elasticity with pinned Code_Aster 15.2. If useOrthotropicBulkProperties is explicitly enabled, Mode-I uses Code_Aster 17.4 and applies one measured homogeneous orthotropic tensor; when explicit layerwise mapping is enabled, each bulk layer uses its G-code-mapped frame while all layers share that same tensor. The mixed-mode Turon route additionally requires same-process-pair DCB, ENF and at least two MMB tests at distinct measured energy fractions, traceable K, and an explicit global displacement with both opening-normal and in-plane tangential components greater than one degree. The displacement's shear axis must match the common measured ENF/MMB shear axis within one degree; reject other directions because this solver law has one tangential response. All ENF and MMB tests must use the same in-plane shear axis within one degree. Supply 1..32 ordered parallel split planes for a same-material layer stack; their normals may be tilted in global coordinates, and the same measured layer law is repeated at every plane. This equivalent-interface assumption does not resolve individual roads or within-layer raster variation. Either route may accept useOrthotropicBulkProperties when the exact-process coupon contains measured E2/E3, nu13/nu23, G12/G13/G23 evidence and a confirmed or traceable global print frame. Check the exact-process coupon match is unambiguous and do not infer axes from a photo or CAD face. Confirm the tested material's side relative to the chosen split-plane normal; do not infer this from the body or face normals. The measured test direction must match the normal/shear mode; the support face lies below the first split plane and the loaded face above the last plane along the shared normal. The server derives peak traction, integrated fracture energy and displacement endpoint from exact tests, exports the selected current Solid, creates a conforming multi-region mesh with a cohesive element set at every requested plane, and aborts if the CAD revision changes. Review the mesh-resolution screen and perform refinement/sensitivity work as needed. Turon approximates measured curves using peak/area parameters and its K still needs sensitivity analysis. The bulk tensor remains one measured single-material continuum whose local frame may vary by mapped layer; repeated cohesive planes use the same measured, direction-independent interface law and do not resolve individual deposited roads, within-layer raster mixtures or direction-dependent adhesion. Use results only as raw solver responses; never report it as a part-strength verdict, qualified layer-adhesion value, print approval or design allowable. Before asking for a physical coupon record, obtain the selected immutable profile hash plus material.nozzleTemperatureC, slicer.layerHeightMm, slicer.nominalInfillPercent, slicer.sparseInfillPattern, slicer.wallLoops, slicer.topShellLayers, and slicer.bottomShellLayers from workbench_manufacturing_profiles when those settings are available. Carry the exact process values into the record; do not infer missing infill or substitute a generic profile. These are resolved profile defaults; sparse fill or shell settings may be overridden per object and must not be assumed when a modifier was used. If the coupon was printed with a different setting, first register/select the matching immutable process profile and hash. When layer positions should follow the actual print, record the measured profile hash and layer height in the physical interface-test and material-coupon process; layer height is part of exact process identity. After slicing, if the job reports a complete deposition-height schedule, call workbench_slicer_interface_heights for every selected interface index. Derive each offset as that interface’s depositionLayerZMm minus firstDepositionLayerZMm and pass the ordered values as interfaceOffsetsMm; this preserves first-layer and adaptive heights. Also pass the selected interface response plus its job/profile/source/G-code hashes, layer count and coordinate frame as depositionPathEvidence so the cohesive report retains the exact per-layer road-orientation observations. This evidence preserves the measured toolpath but does not qualify material properties. Use workbench_slicer_layer_path_orientations in batches of up to 32 to retrieve every layer direction (up to 33 total layers) when the user wants layerwise solver orientation. Confirm how slicer X maps into the CAD global frame and that the exact-process coupon axis 1 represents the dominant deposited-road direction; pass those confirmations as pathFrameMapping and roadAxisMapping. Combine complete responses, including the final deposited layer, as layerPathEvidence with shared job/profile/source/G-code hashes. Every layer must have complete linear or planar circular-arc coverage; do not treat unsupported arc/spline moves as complete. With explicit roadAxisMapping and useOrthotropicBulkProperties, Mode-I and Turon use one measured tensor with a separate local frame per layer. Without roadAxisMapping, frames remain candidates and do not affect solver response. This does not model multiple materials, layer-varying properties, within-layer raster mixtures or directional interface adhesion. Returned coordinates are in slicer build coordinates: map only relative offsets onto the confirmed CAD print axis, never copy absolute slicer Z into CAD. Otherwise the planner uses the nominal profile height. Call plasticity_plan_cohesive_layer_planes with that same profile hash and layer height, a point on the first interlayer plane after object placement, confirmed global build direction, total layer count and explicit interface indices. Copy both its plan and planes into layerPlanePlan and splitPlanes; analysis rejects legacy test records without a recorded layer height and any height that differs from the measured process profile. It also checks profile identity, plane coordinates, and (for orthotropic bulk) alignment with the coupon’s confirmed build direction. For up to 32 interfaces the planner requires the complete stack; for larger stacks it marks selected planes as incomplete, and omitted interfaces remain unanalyzed. Do not present selected-plane analysis as full-stack delamination resistance. If no validated placement/anchor is available, ask instead of inferring layer positions from an image or display mesh. Read maximumVonMisesElementSICN alongside each raw stress-peak locator to see the Gmsh SICN of that exact tetrahedron. Compare it with the mesh minimum only as local mesh-shape context; neither a high value nor separation from the minimum proves stress accuracy, a resolved hotspot, convergence, or strength. Read maximumPrincipalStressMPa and minimumPrincipalStressMPa as raw tensile/compressive principal extrema across sampled integration points, each with its own mesh locator. Their refinement trends and signed relative changes expose mesh sensitivity only; they are diagnostic and must never be compared with a von Mises allowable or reported as a pass without a criterion qualified for the selected failure mode. The public FEA tool measures the rank of all fixed global translations at the actual mapped mesh nodes and stops before CalculiX if they leave any of the six rigid-body translation/rotation modes unconstrained. A full rank of six is necessary to remove those rigid-body modes, but it does not establish physical support validity, elastic stability, or absence of local mechanisms. For a heat-set insert, use pullout and torque-out capacity only when the evidence matches the exact insert, host material, print profile, orientation, pocket and installation process. Ask for the worst-case demand on one insert; do not divide a group load evenly without a load-path model. For two or more fasteners under an in-plane load, establish every transfer-point coordinate, both force components, the point of application and any free moment. Use the elastic group method only after confirming a rigid attachment and identical in-plane fastener stiffness. Use each fastener’s own vector resultant for any member check. On a rectangular mounting face, call plasticity_check_fastener_group_layout with an explicit opposedFaceId to verify matching native perforated faces and exact plate thickness. This is geometry evidence only and does not establish a load path or capacity. plasticity_verify_fastener_group_plate_bearing compares each elastic per-fastener demand with a directly traceable, configuration-matched, already factored bearing allowable using exact measured thickness and hole diameter. It can additionally check a straight transverse net-tension section only when you provide the external tensile resultant separately, identify local X or Y as its axis, supply a distinct traceable factored tensile allowable, and confirm uniform membrane tension, centered through-thickness loading and a straight transverse failure path. Never substitute per-fastener demands for the external plate tension or infer that resultant from a sketch. Optionally use edgeShearOut with a separate traceable, factored shear allowable to check local two-plane tear-out for per-fastener vectors aligned to local X or Y; diagonal vectors and e/d below 1.5 are unsupported, while e/d below 2 remains conditional. Angled/staggered fracture paths, compression-side buckling, shared-ligament interaction, unsupported tear-out directions, bypass and complete-joint strength remain unchecked; even a within-allowable result is only a conditional local screen. If a physical multi-hole plate test is run, record the measured specimen, exact hole layout, print-process/profile, fixture, load axis, individual peak loads and observed failure modes with plasticity_record_fastener_group_test, then use plasticity_match_fastener_group_test to find only an exact configuration match. A test match is evidence only: it does not produce a design allowable or pass a different part or support setup. To add a test benchmark to the CAD-bound report, first confirm that the record's exact process and fixture/load path apply to the current part. Supply the selected immutable record ID, an independently evidenced dimensional equivalence tolerance, the exact process, safety factor, and explicit process/fixture confirmations; the server then checks every measured plate and hole dimension against the live B-rep. This comparison only checks factored external tensile demand against the lowest observed specimen peak. It is not a statistically reduced allowable, strength pass/fail, or proof for unobserved failure modes. Do not select a nearby test by appearance or silently assume fixture equivalence. The single-through-fastener plate method assumes one hole; never repeat it per hole and aggregate the results into a multi-hole plate or whole-joint pass. Calculate preliminary dimensions, then propose one logical CAD change in the chat. Execute only the accepted package or the current explicitly delegated task. Delegation ends when the task ends and is not restored after restart. Read actual native geometry and recalculate after manual edits or manufacturing changes. Workbench is optional. Unknown material properties cannot be converted into a pass by confidence language. For an end-to-end tongue-root scenario, use plasticity_calculate_tongue_root_strength_from_coupon_data or plasticity_verify_tongue_root_strength_from_coupon_data. They match the physical coupon record to the exact Workbench profile hash, printer, material, orientation, infill, nozzle temperature and measured slicer layer height, then copy only measured Young's/shear moduli and their evidence into the calculation. Supply separately sourced tensile/shear design allowables and explain their applicability; raw coupon strengths are never substituted for allowables. These tools keep material suitability unconfirmed and the result conditional. No-match or conflicting records return without a calculation. If no record exists, ask for the report and unresolved process details, and record it only after the user confirms the physical tests. The registry does not check test-standard compliance, derive statistical design values, or certify a part. Never substitute a generic filament label or slicer properties for coupon evidence. For axial rectangular or rectangular-cantilever scenarios with exact-process coupon data, use plasticity_calculate_rectangular_strength_from_coupon_data. It copies only measured Young's modulus and its evidence. Provide independent tensile design allowable evidence, plus an independent compressive allowable for a cantilever; state their applicability basis. Never map raw coupon tensile/compressive strengths into design allowables. These scenarios remain conditional with material suitability unconfirmed. Exact-process no-match or ambiguity returns without saving a calculation. Nominal section stress is not whole-part validation.

Use the bounded, action-free Codex API profile to inspect the supplied photo/sketch views and text (up to four PNG/JPEG/HEIC/HEIF images, 20 MiB each; macOS converts HEIC/HEIF locally to JPEG for analysis). Returns structured functional interfaces and candidate features, an explicit scale-confidence status, and at most one next decision-relevant question package. A dimensioned/calibrated result requires traceable positive millimeter scale evidence; unscaled views cannot yield millimeter measurements. It never creates CAD or sends a print. The request is durable and is not retried under the same ID after interruption.

Read a persisted design-reference analysis without starting or retrying a Codex turn.

Read one persisted strength-analysis request without starting or retrying Codex.

Calculate and persist a labelled scenario from caller-supplied dimensions. A supplied CAD binding is not treated as verified geometry.

Calculate axial-tension or rectangular-cantilever scenarios using Young's modulus copied from an exact printer/material/profile/orientation coupon record. Tensile allowable, and for cantilevers compressive allowable, must be separately evidenced. Raw coupon strengths are never mapped to allowables. Process suitability remains unconfirmed; no-match or ambiguity returns without saving a calculation.

Evaluate an explicit finite list of rectangular-member heights or plate thicknesses. This is a deterministic scenario calculation and does not modify CAD.

Read exact native B-rep topology and verify one constant rectangular Solid without using display or mesh bounds.

Measure one exact rectangular panel wall from opposed native planar faces of the same Solid. Verifies face outlines, opposite normals, coplanar alignment and wall thickness; no display or mesh bounds are used.

Read one exact current planar Solid face from native B-rep line and circular boundaries. Display bounds and render meshes are not measurement evidence.

Read the exact native B-rep intersection of one Solid with an arbitrary plane. The temporary Sheet and cut results never enter the persistent document.

Inspect a caller-selected exact native section and derive root width/thickness only when the B-rep boundary is one axis-aligned rectangle with four straight edges and no holes. The returned plane binding can be used by plasticity_verify_tongue_root_strength.

Inspect 1–32 explicitly supplied candidate planes through one Solid. Every exact native section must belong to the same current Plasticity session, document, body and revision; individual planes may be unsupported if they do not cut a supported section.

Sample 2–32 exact B-rep sections at evenly spaced stations along an explicit plane normal and offset interval. The caller chooses the scan direction and range; this does not identify or rank mechanically critical sections.

Measure exact B-rep sections and calculate one identical load/material scenario at each explicitly bounded station. It ranks only supported, complete single-mode utilization results; any unsupported or incomplete station suppresses a global governing-station claim. It saves one immutable scan report with every measured input and calculation; report reads re-check the live CAD binding.

Read one immutable section-strength scan, including exact candidate geometry, evidence, calculations and ranking. Re-checks the stored session/document/revision/body against the live Plasticity binding and reports current, stale or unverified.

Verify one exact constant-thickness rectangular Solid plate with one cylindrical through-hole. The load direction selects the loaded edge; no display or mesh bounds are used.

Calculate and persist a labelled planar-section scenario. Caller-supplied geometry remains unverified and any supplied CAD binding is removed.

Calculate and persist bearing, loaded-edge shear-out and net-section tension for one caller-described through fastener. Caller CAD bindings are removed.

Calculate and persist a screening scenario for one fastener in axial tension, direct shear, and NASA combined tension-shear interaction. Effective areas, plane count, plane location, loads, and material limits must be explicit and traceable.

Calculate and persist a root-only screening scenario for a rectangular tongue under transverse point load. Requires traceable geometry, load, orientation-matched effective material properties, allowables, shear correction factor, safety factor, and deflection limit. A result never validates the complete tongue-and-groove joint.

Resolve one exact physical coupon record by printer/material/profile hash/orientation/infill percentage and pattern/wall loops/top-bottom shell layers/temperature, copy only its measured Young's and shear moduli plus source evidence, and calculate a labelled tongue-root scenario. Tensile and shear design allowables must still be supplied with independent evidence and an applicability basis. No-match or ambiguity returns without saving a calculation; material suitability remains unconfirmed, so this cannot produce a pass.

Re-read the bound exact native section, replace root width and thickness with measured B-rep dimensions, calculate and persist a CAD-bound root-only screening report, then recheck the live section before saving. This does not validate the complete tongue-and-groove joint.

Use an exact process-matched physical coupon record for the measured moduli, keep separately evidenced design allowables unconfirmed, re-read the bound arbitrary-plane native section, replace root dimensions with exact B-rep measurements, and recheck the same session/document/revision/topology before saving. No-match or ambiguity returns without modifying CAD or saving a calculation. A result never validates the complete tongue-and-groove joint.

Calculate and persist axial stripping and tensile-failure checks for one tapped hole, nut, or threaded insert. All three allowable loads, engagement, fully formed thread count, load, and evidence must be explicit; nominal M size alone never supplies capacity.

Calculate and persist independent pullout and torque-out screening for one installed heat-set insert. Qualification capacities must match the insert, host, print, pocket, and installation process; simultaneous modes remain conditional.

Calculate and persist elastic in-plane force and moment distribution for a caller-described rigid group of identical-stiffness fasteners. Any supplied CAD binding is removed. Loads are distributed exactly as supplied with no safety factor. Optionally provide a traceable design allowable in N for every exact fastener configuration in shearCapacities; each allowable must already include its required safety factor. This adds an individual fastener shear-only screen, not a joint or plate strength pass. Without these records, the tool reports load demand only.

Re-read exact cylindrical faces from one current Solid, replace caller fastener coordinates with native B-Rep axis centers, calculate the in-plane load distribution and persist a CAD-bound report. Optionally compare each demand with a traceable configuration-matched shear design allowable that already includes the required safety factor. The optional screen covers fastener shear only, never joined-plate or whole-joint strength. It does not mutate CAD.

Re-read the exact front and opposed native B-rep faces and cylindrical hole faces of one rectangular multi-hole plate, compare each fastener's elastic in-plane demand with a traceable factored bearing allowable, and optionally check straight transverse net tension or local two-plane edge shear-out. A physical-test benchmark requires an immutable record ID, exact print process, an evidence-backed dimensional equivalence tolerance, and explicit confirmation that process and fixture/load path match. Net tension uses an explicitly supplied external tensile resultant along local X or Y and the minimum straight cut across measured circular holes. Local edge shear-out uses each fastener's elastic resultant only when it aligns with a local rectangle axis; diagonal demands and e/d below 1.5 are unsupported. Supply separate traceable factored tensile and shear allowables and confirm the method assumptions. Angled/staggered fracture paths, compression, shared-ligament interaction, bypass and the complete joint remain unchecked; no overall pass is returned.

Read an immutable multi-hole local-bearing, optional straight-cut net-tension and edge shear-out report and re-check its group and exact native plate geometry binding against the live Plasticity document.

Verify exact current rectangular B-rep dimensions for a member or separate plate body, append measured evidence, recalculate and persist a CAD-bound report. It does not mutate CAD. Search accessible primary product/material sources before asking the user for known facts. Before requesting missing print-material data, call plasticity_plan_single_material_strength_tests for the selected method and exact one-material process. Use its measurement matrix to ask only for evidence needed by that route; explain why the measurements matter and never invent coupon values or allowables. A DCB task may contain an explicitly labeled generic-PLA literature geometry as a starting reference only; do not present it as a Creality property, normative specimen size, or sample-count requirement, and require checking the selected process, fixture and method. Keep the confirmed road/build axes and same-material layer-interface assumptions explicit. For Creality PLA literature references, read plasticity://strength/interlayer-literature-baseline. It contains separate CR-PLA and Hyper PLA records plus generic-PLA Z-tension/DCB context; do not merge product families or transfer values across SKUs. It is a screening reference only: do not register literature values as physical coupons or interface tests, bind them to the user's exact K1C process, treat them as design allowables, or infer a traction-separation curve from fracture energy. Manufacturer-mirror discrepancies and non-matching printer/process tests must remain visible; ask for exact-process physical tests before a calibrated cohesive result. When the user provides raw direct-tension machine CSV, call plasticity_import_interface_tensile_csv first with explicit specimen-ID/force headers, delimiter, decimal separator, force unit and sign, measured net area and observed failure location for every coupon. Review its hashed peak-force preview; the importer does not filter or correct machine data, register a test, or derive DCB/ENF/MMB curves. For directly loaded specimens supplied in another format, record every sample's measured peak force, net cross-section, observed failure location and SHA-256/source locator, then call plasticity_calculate_interface_specimen_strengths; report force/area as nominal specimen stress. Include only confirmed interface failures in its summary statistics; show bulk/fixture failures individually and do not pool them with interface failures. Keep all statistics descriptive only. Do not call these local interface tractions or design allowables, and do not use them as a cohesive law. When a raw DCB force/displacement CSV is available, use plasticity_import_dcb_mode_i_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting explicitly, then select the exact source record for every manually observed crack-growth point and enter that measured crack length; never let the importer filter traces, infer growth or select peak loads. It converts units only, preserves the source hash/record locators and calculates an exploratory MBT G_I(a) curve. Require the caller to establish machine-compliance-corrected load-point displacement and quasi-static linear-elastic behavior; attestations are not independent verification. Review selected rows, crack lengths, calculation and limitations against the physical log. If no CSV is available, plasticity_calculate_dcb_mode_i_energy accepts equivalent manually selected observations with traceable source hash/locators. This does not conform the test to ASTM D5528 (whose stated scope is unidirectional fiber composites), create a traction-separation law, or qualify CR-PLA. Never convert G_I into peak traction or strength without separate evidence. When raw ENF force/displacement CSV is available, use plasticity_import_enf_mode_ii_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting; group calibration rows into each measured run/crack length, manually select only the initial-linear records, and manually select the observed initiation/peak record for the fracture run. The tool fits compliance from the selected rows, then the existing ENF calculator fits C against a^3 and returns exploratory G_IIc; it never searches for the linear range, crack growth or peak. Review each compliance fit, source hash/record locator, force/displacement correction and fixture log. If pre-reduced compliances are supplied, use plasticity_calculate_enf_mode_ii_energy directly. Neither path determines ASTM D7905 validity for printed PLA, registers physical evidence, builds an R-curve or yields a cohesive law. Keep the confirmed in-plane shear direction exact; do not merge ENF runs with differing direction just because their layer normals match. When the requested failure mode is layer separation, distinguish CR-PLA's physical tensile/infill study and the two-sample CR-PLA-associated vertical layer-adhesion screen from the Hyper PLA flat tensile/flexural study: observed flexural delamination is qualitative failure-mode evidence, not a measured interface law. The baseline also records an upright generic-PLA tensile coupon on a Creality Ender 3 Pro and a custom generic-PLA interface coupon/calibrated cohesive parameter; neither identifies the filament as Creality nor establishes a same-process cohesive calibration. For only a nominal normal-strength screen across the layers, use the focused layer-interface-normal-tension scope; require measured failure at the interface and do not use that peak as a cohesive law. Use layer-interface-mode-i for a Mode-I fracture response, layer-interface-mode-ii for exploratory ENF Mode-II initiation energy using a same-fixture compliance calibration, and layer-interface-mixed-mode only when the requested analysis needs DCB/ENF/MMB cohesive evidence. The focused ENF calculator fits the experimentally supplied compliance against crack length cubed and uses the initiation peak; it is not an R-curve, does not verify the physical fixture or raw compliance fit, and does not claim ASTM D7905 validity for printed PLA. Layerwise static FEA assumes perfectly bonded interfaces and cannot answer a delamination question. After reviewing an exploratory DCB energy preview against the physical test log, persist it only through plasticity_record_dcb_mode_i_energy_test with explicit confirmation that the measurements came from real physical tests. Retrieve a full record with plasticity_read_dcb_mode_i_energy_test and require exact process/interface-normal/protocol matching through plasticity_match_dcb_mode_i_energy_test before comparing runs. For ENF Mode-II energy, record reviewed measurements only through plasticity_record_enf_mode_ii_energy_test with the same physical-test confirmation; retrieve all calibration/fracture inputs with plasticity_read_enf_mode_ii_energy_test and exact-match process/interface-normal/in-plane-shear-axis/protocol using plasticity_match_enf_mode_ii_energy_test. Both immutable energy registries are separate from peak-strength and traction-separation records and are not consumed by cohesive FEA. For an exploratory mixed-mode initiation partition, use plasticity_calculate_mmb_mode_i_ii_energy only when measured MMB force and geometry, same-process flexural and orthotropic moduli, and material-axis mapping are available; require lever weight to be measured negligible or counterbalanced. This Reeder-Crews beam-theory estimate does not establish ASTM D6671 validity for printed PLA and does not replace full mixed-mode traction-separation curves or provide a cohesive law. For raw MMB force CSV, use plasticity_import_mmb_mode_i_ii_energy_csv only with manually selected initiation records and a physically observed criterion; it does not search traces for onset/peak. Record a preview only after reviewing the source and confirming the data are from physical tests; matching caller confirmation with plasticity_record_mmb_mode_i_ii_energy_test persists and server-recomputes this separate energy estimate. Use plasticity_match_mmb_mode_i_ii_energy_test only for the exact process, interface normal, shear axis and protocol; read the full evidence with plasticity_read_mmb_mode_i_ii_energy_test. This registry is not cohesive input or an FEA source. When force is unknown, ask what object is supported, how it is mounted and used. Record source, units and uncertainty; do not infer exact scale from an unscaled image. Ask at most one next-step question package per response, then wait for the user's answer. Include only facts needed to choose the next safe step; defer material, manufacturing, tolerances and detailed dimensions until they affect that decision. Re-evaluate after every answer and ask a focused follow-up only when it changes the method, required evidence or next action. If the user does not know, move to one useful contextual clue such as the supported object, use, environment or mounting; do not repeat a list of unknowns or guess. On a new bracket task, first ask compactly what it supports, its load/use and how it is mounted; defer section, material/process and displacement questions until that first answer narrows the load path and method. Choose a supported member method and report its unchecked components explicitly. For a flat rectangular panel, establish net pressure and the real condition of all four edges before choosing the simply-supported plate method. Do not infer edge support from appearance. For a straight prismatic rectangular member in centred axial compression, use the Euler column method only when effective-length factor K, elastic limit, compressive allowable, and the actual restraint condition are supported by evidence. Use the weakest section axis, require a negative force value for compression, and reject Euler results when its predicted critical stress exceeds the supplied elastic limit; do not guess K or treat the method as an inelastic, eccentric-load, local-buckling, or whole-part check. For an integral rectangular enclosure wall, inspect two opposed planar faces on the same Solid with plasticity_inspect_integral_rectangular_plate, then use plasticity_verify_integral_plate_strength to replace dimensions from exact native B-rep. The inspector accepts only rectangular faces whose outlines match or inset by one wall thickness on each side. It measures geometry only: establish the real support, pressure, material and edge conditions separately; never assume a box wall is simply supported. For section analysis, identify the critical plane and explain the load path that makes it critical. Use an existing planar face when it matches; otherwise inspect an arbitrary plane through the Solid. For one fastener carrying in-plane plate load, establish the load direction and separate bearing, shear and tensile allowables before calculating. Never substitute compressive strength for bearing strength. For a fastener group on a rectangular planar face, inspect exact hole centers and boundaries, then check center-to-edge distance, hole-edge clearance, pitch, ligament and every applicable head, washer, nut or driver envelope against explicitly sourced or user-approved criteria. A measured layout without criteria is not a pass, and a layout pass is not a strength pass. When checking the fastener member, establish its grade, tensile stress area, effective shear area at the actual plane, one or two shear planes, total axial tension including applicable preload, and actual shear load. Explain why the selected combined-load interaction is applicable before accepting it. For a tapped hole, nut, or threaded insert under axial load, establish the designation, pitch, actual engagement, fully formed engaged thread count, and configuration-matched allowable loads for internal-thread stripping, external-thread stripping, and fastener tension. Do not infer any capacity from nominal M size. A procured nut or insert requires a specified assembly allowable or dedicated test rather than an unqualified nominal shear-area calculation. For solver-backed static FEA, first use plasticity_match_material_coupon_data when a physically tested material record may match the print process. Only an unambiguous exact match for printer, material, profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature and measured slicer layer height can supply Young's modulus; pass the selected record ID/process and the exact recorded modulus so the server binds and validates it. If match is ambiguous because compatible partial records have no single consolidated record, ask the user for the confirmed count of unique physical specimens and call plasticity_combine_material_coupon_data with those exact record IDs; it preserves measured values and evidence without averaging. If records conflict, stop and ask the user which physical measurement applies. For an orthotropic print model, do not use a uniaxial coupon as a full tensor. When an immutable exact-process record contains the complete measured tensor and nu12, pass orthotropicMaterial.couponRecordId and process, plus materialCoupon with the same record ID; the MCP checks E1, nu12, every remaining constant, each evidence object and the confirmed global print axes, and rechecks the record when reading the saved report. Otherwise pass E2/E3, nu13/nu23, G12/G13/G23 with separate exact evidence, global material axes 1 and 2, and a user-confirmed or sourced global build direction; material axis 3 must align with that build direction and represents the homogeneous layer-normal response. This does not represent individual layers or prove adhesion. Model one material and one print process per analyzed part; do not combine material datasets or infer a multi-material print. When the selected exact-process coupon record contains a qualified Tsai-Wu dataset, pass its ID as orthotropicMaterial.tsaiWuQualificationRecordId along with the identical orthotropicMaterial.process; the server loads measured strengths, biaxial interaction data and source evidence, then verifies the axes. Do not copy or reconstruct criterion values by hand. If the exact-process match remains ambiguous after consolidation, ask the user to resolve conflicting measurements; never choose a record silently. The legacy youngsModulusMPa and poissonRatio fields are E1 and nu12. Ask the user to resolve the print-axis orientation if it is unknown; never infer it from a photo or CAD face. The server checks positive-definite compliance and rejects a von Mises allowable or coupon binding for this orthotropic model. Otherwise attach youngsModulusEvidence whose value exactly matches the modulus; measured/sourced evidence needs a URL, SHA-256 and locator. poissonRatioEvidence is always required and must exactly match the ratio. A physical coupon record may store measured nu12 with its exact process. If using its Tsai-Wu record-binding path, the server verifies the nu12 value and evidence against that record; otherwise supply the input evidence directly from a traceable source. Never infer nu12 from a generic plastic label. An assumed value must include its reason and stay explicitly scenario-only after discussing the assumption with the user. Optionally supply a directly measured or sourced, process-applicable factored von Mises design allowable using factoredVonMisesAllowableMPa, exact matching factoredVonMisesAllowableEvidence with URL, SHA-256 and locator, and factoredVonMisesAllowableBasis. It must already include the design factors. Never turn a generic tensile strength or raw coupon peak into a design allowable. The report will compare every sampled raw mesh peak with it as a diagnostic screen only: above means a sampled peak exceeds that supplied allowable; below does not prove strength. This never changes strengthPass or print approval. Confirm the actual restraints with the user. Legacy supportFaceIds fix all three global translations on each selected planar face; use explicit supportConditions only when the user has specified which global x/y/z translations are zero on each face. Each condition acts on all nodes of that face, is not a frictionless-contact or rotational support, and may leave rigid-body modes or create a singular system; do not infer it from a photo or face normal. Supports and loaded faces must not share mesh nodes. Every selected planar load face needs an explicit uniform traction and/or resultant force with its point of application and free moment. Use plasticity_analyze_static_fem only for one Solid with 1–8 explicit support conditions and load vectors grounded in a selected planar face. Use named loadCases for physically distinct scenarios such as weight, operating force and handling load; do not combine mutually exclusive scenarios. Each case has a separate solver result, and comparison proceeds only when the generated meshes are byte-identical. Set meshRefinementSteps to 1–3 for two to four mesh levels when a trend across successively halved element sizes is useful; the report classifies the sampled direction of raw maximum von Mises stress and observed displacement as increasing, decreasing, unchanged, non-monotonic or insufficient-levels. For more than four levels, run separate analyses against the same current CAD revision and call plasticity_compare_static_fem_refinement_reports; it merges only reports with matching loads, supports, material evidence, native geometry and byte-identical overlapping mesh results. Check the returned freshness before relying on it. Treat all trend labels and relative changes as diagnostic evidence only. Never call a mesh trend a pass or proof of convergence. Each result includes the raw maximum C3D4 integration-point stress element and its mesh-element centroid in millimetres; this is a mesh-bound locator, not an averaged stress field, a resolved critical-region boundary, or proof of a physical hotspot. Locations may move between refinement levels. The legacy top-level load fields still represent one case. The tool returns total and per-support reaction forces and moments, global force/moment equilibrium residuals and a named displacement axis for each case; per-support values are diagnostic resultants over each support node set. Re-read it with plasticity_static_fem_report; any CAD revision change makes it stale.

For layerwise static FEA, keep the single-material exact process consistent from measurement through solver input. Read the immutable profile hash, infill pattern, wall loops, top/bottom shell layers and measured layer height from workbench_manufacturing_profiles, then slice the intended model/orientation with the selected Creality K1C profile. Record actual specimen settings if per-object overrides differ from profile defaults; an unchanged profile hash does not erase those differences. Call workbench_slicer_layer_path_orientations in batches of up to 32 and combine layer-path results for every deposited layer, including the final deposited layer, into layerPathEvidence with identical job ID, profile hash, source-artifact hash and G-code hash. Layerwise static orientation is supported only for a complete linear or planar circular-arc direction result for every layer in stacks of 2–33 layers; do not fill missing or curved-path directions by inference. Confirm pathFrameMapping.slicerXDirectionGlobal in CAD global coordinates and explicitly confirm that the same exact-process coupon's material axis 1 represents the dominant deposited-road direction in roadAxisMapping. Call plasticity_plan_cohesive_layer_planes with the same profile hash and layer height, current CAD anchor at the first interlayer plane, confirmed CAD build direction, total layer count, and the full layer-path evidence/mappings; pass its returned plan as layerPlanePlan to plasticity_analyze_static_fem. When the slicer supplies actual interface heights, call workbench_slicer_interface_heights for every interface in this complete stack, set each interfaceOffsetsMm value to its depositionLayerZMm minus firstDepositionLayerZMm, and attach the matching job/profile/source/G-code hashes as depositionPathEvidence; this preserves first-layer and adaptive heights instead of assuming nominal uniform spacing. The static analyzer requires the plan to match the one measured orthotropic process and applies the same measured orthotropic tensor to each layer with its confirmed G-code-mapped frame. It assumes perfectly bonded layers and cannot assess delamination. Do not use static FEA to assess delamination; use the separate same-material cohesive route only when matching physical interface tests are available. Never infer layer directions, material identity or interlayer strength from a photo, generic material label or nominal slicer preset.

For orthotropic FEA, optionally supply all nine directly traceable, already factored X/Y/Z tensile and compressive plus XY/XZ/YZ shear limits in orthotropicMaterial.factoredAllowables, with separate exact-value evidence and an applicability/design-factor basis. Never substitute generic datasheet strength or an unqualified coupon peak. The returned componentwise maximum-stress screen uses local material-axis stress extrema and is diagnostic only; it assumes one homogeneous orthotropic continuum, does not model layer interfaces, delamination or different-material joints, and omits multiaxial interaction. Matched interlayer-test data can inform Z-tension and XZ/YZ-shear allowables but does not turn this into a cohesive-interface analysis. Never report this screen as verified layer adhesion, part strength or print approval. For an optional 3D Tsai-Wu first-failure screen in orthotropic linear FEA, require one exact single-material printer/material/profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers and nozzle-temperature identity in orthotropicMaterial.process. Prefer binding the unique exact-process qualification by its orthotropicMaterial.tsaiWuQualificationRecordId; the server loads its nine directly measured, un-factored X/Y/Z tensile/compressive and XY/XZ/YZ shear failure strengths, three normalized XY/XZ/YZ normal-interaction coefficients and evidence, then verifies material axes. Each strength test must attest its material-frame axis and mode (for example, x tension is material-1 tension; xy shear is material-1-2 shear); each interaction and its source dependencies must attest the corresponding biaxial plane. Legacy records without these direction attestations remain readable but cannot qualify a Tsai-Wu FEA. A complete inline orthotropicMaterial.tsaiWuCriterion remains supported when needed. Interactions must be derived from traceable biaxial tests. Ask the user for these records if missing; never copy generic datasheet values, assume the conventional interaction coefficient or infer it from uniaxial coupons. The normalized interaction matrix must be positive definite. The result reports local integration-point failure indices and proportional load factors to index one only. A load factor is not a design safety factor, and neither a sub-unity index nor a large load factor means the part passed. The model still represents one homogeneous material, not individual roads or delamination.

When actual test-coupon G-code is available, preserve its profile/source/G-code hashes, selected per-layer road summaries and explicitly user-confirmed slicer-to-global axes in depositionPathEvidence. This is test provenance only, not an adhesion measurement or solver input; never reconstruct road direction from nominal slicer settings.

When physical adhesion between printed layers of one material is relevant, use plasticity_record_material_interface_test only for caller-provided measured test results with the same exact printer/material/profile process on both sides, interface normal, test mode, load direction, fixture/specimen protocol hash and observed failure location. Normal-tension load direction must align with the interface normal; interface-shear load direction must lie in its plane; mixed-mode must contain both components. Store full compliance-corrected pure-mode curves as scalar separation/traction data and MMB curves as separate normal/tangential separation and traction components, with source SHA-256 and locator. Call plasticity_analyze_material_interface_test_curve to summarize measured work and mode mixity. When the user has a CSV of already processed physical fracture data, use plasticity_import_interface_fracture_csv for a read-only per-specimen preview; explicitly map the method, columns, units and CSV formatting, and attest that the values are already compliance-corrected physical traction-separation data. Never convert raw machine force-displacement data with this importer. Verify each source hash/record locator, specimen, fixture, exact print process and observed failure plane before separately recording any curve with plasticity_record_material_interface_test. For a CZM_TURON candidate fit, provide Mode-I DCB, Mode-II ENF and at least two MMB tests at distinct measured energy fractions; all ENF and MMB tests must use the same in-plane shear axis within one degree because this route has a single tangential cohesive law. The MCP requires those protocol identifiers in each testMethod and reports the pure-mode peaks plus fit residuals. Review residuals against test uncertainty. Do not invent ETA_BK. K is not identified by that fit and must not be silently guessed. Code_Aster CZM_TURON uses one normal and one tangential cohesive response, sharing tangential strength and fracture energy across both in-plane tangent directions; it cannot represent direction-dependent shear adhesion. Matching shear axes prevents mixing directional datasets but does not prove that the bond is isotropic. These outputs summarize physical evidence only; they are not qualified cohesive-law parameters, design allowables or a part FEA. Do not claim bond integrity from the homogeneous orthotropic FEA screen. Use plasticity_analyze_cohesive_interface only for a single-material printed part, with an exact-process coupon for that one bulk material, traceable Poisson ratio and a matching immutable same-material layer test. Dissimilar-material printed bonds are rejected before meshing and are outside this calculation scope. Mode-I analysis requires a full normal-tension DCB traction-separation curve with failure at the layer interface. Its modeILaw may explicitly select CZM_EXP_REG or CZM_LIN_REG; omitted input preserves the CZM_EXP_REG default. Both parameterize their softening law from measured peak traction and integrated fracture energy and do not fit the full measured curve shape. Review this choice against the measured curve and record the reason; do not infer it from part geometry. Read the returned solver.result.v3Interpretation: for CZM_EXP_REG, V3 is a damage variable in [0,1]; for CZM_LIN_REG, V3=2 means the cohesive element is completely broken, so do not present it as the same normalized damage fraction. By default it uses isotropic bulk elasticity with pinned Code_Aster 15.2. If useOrthotropicBulkProperties is explicitly enabled, Mode-I uses Code_Aster 17.4 and applies one measured homogeneous orthotropic tensor; when explicit layerwise mapping is enabled, each bulk layer uses its G-code-mapped frame while all layers share that same tensor. The mixed-mode Turon route additionally requires same-process-pair DCB, ENF and at least two MMB tests at distinct measured energy fractions, traceable K, and an explicit global displacement with both opening-normal and in-plane tangential components greater than one degree. The displacement's shear axis must match the common measured ENF/MMB shear axis within one degree; reject other directions because this solver law has one tangential response. All ENF and MMB tests must use the same in-plane shear axis within one degree. Supply 1..32 ordered parallel split planes for a same-material layer stack; their normals may be tilted in global coordinates, and the same measured layer law is repeated at every plane. This equivalent-interface assumption does not resolve individual roads or within-layer raster variation. Either route may accept useOrthotropicBulkProperties when the exact-process coupon contains measured E2/E3, nu13/nu23, G12/G13/G23 evidence and a confirmed or traceable global print frame. Check the exact-process coupon match is unambiguous and do not infer axes from a photo or CAD face. Confirm the tested material's side relative to the chosen split-plane normal; do not infer this from the body or face normals. The measured test direction must match the normal/shear mode; the support face lies below the first split plane and the loaded face above the last plane along the shared normal. The server derives peak traction, integrated fracture energy and displacement endpoint from exact tests, exports the selected current Solid, creates a conforming multi-region mesh with a cohesive element set at every requested plane, and aborts if the CAD revision changes. Review the mesh-resolution screen and perform refinement/sensitivity work as needed. Turon approximates measured curves using peak/area parameters and its K still needs sensitivity analysis. The bulk tensor remains one measured single-material continuum whose local frame may vary by mapped layer; repeated cohesive planes use the same measured, direction-independent interface law and do not resolve individual deposited roads, within-layer raster mixtures or direction-dependent adhesion. Use results only as raw solver responses; never report it as a part-strength verdict, qualified layer-adhesion value, print approval or design allowable. Before asking for a physical coupon record, obtain the selected immutable profile hash plus material.nozzleTemperatureC, slicer.layerHeightMm, slicer.nominalInfillPercent, slicer.sparseInfillPattern, slicer.wallLoops, slicer.topShellLayers, and slicer.bottomShellLayers from workbench_manufacturing_profiles when those settings are available. Carry the exact process values into the record; do not infer missing infill or substitute a generic profile. These are resolved profile defaults; sparse fill or shell settings may be overridden per object and must not be assumed when a modifier was used. If the coupon was printed with a different setting, first register/select the matching immutable process profile and hash. When layer positions should follow the actual print, record the measured profile hash and layer height in the physical interface-test and material-coupon process; layer height is part of exact process identity. After slicing, if the job reports a complete deposition-height schedule, call workbench_slicer_interface_heights for every selected interface index. Derive each offset as that interface’s depositionLayerZMm minus firstDepositionLayerZMm and pass the ordered values as interfaceOffsetsMm; this preserves first-layer and adaptive heights. Also pass the selected interface response plus its job/profile/source/G-code hashes, layer count and coordinate frame as depositionPathEvidence so the cohesive report retains the exact per-layer road-orientation observations. This evidence preserves the measured toolpath but does not qualify material properties. Use workbench_slicer_layer_path_orientations in batches of up to 32 to retrieve every layer direction (up to 33 total layers) when the user wants layerwise solver orientation. Confirm how slicer X maps into the CAD global frame and that the exact-process coupon axis 1 represents the dominant deposited-road direction; pass those confirmations as pathFrameMapping and roadAxisMapping. Combine complete responses, including the final deposited layer, as layerPathEvidence with shared job/profile/source/G-code hashes. Every layer must have complete linear or planar circular-arc coverage; do not treat unsupported arc/spline moves as complete. With explicit roadAxisMapping and useOrthotropicBulkProperties, Mode-I and Turon use one measured tensor with a separate local frame per layer. Without roadAxisMapping, frames remain candidates and do not affect solver response. This does not model multiple materials, layer-varying properties, within-layer raster mixtures or directional interface adhesion. Returned coordinates are in slicer build coordinates: map only relative offsets onto the confirmed CAD print axis, never copy absolute slicer Z into CAD. Otherwise the planner uses the nominal profile height. Call plasticity_plan_cohesive_layer_planes with that same profile hash and layer height, a point on the first interlayer plane after object placement, confirmed global build direction, total layer count and explicit interface indices. Copy both its plan and planes into layerPlanePlan and splitPlanes; analysis rejects legacy test records without a recorded layer height and any height that differs from the measured process profile. It also checks profile identity, plane coordinates, and (for orthotropic bulk) alignment with the coupon’s confirmed build direction. For up to 32 interfaces the planner requires the complete stack; for larger stacks it marks selected planes as incomplete, and omitted interfaces remain unanalyzed. Do not present selected-plane analysis as full-stack delamination resistance. If no validated placement/anchor is available, ask instead of inferring layer positions from an image or display mesh. Read maximumVonMisesElementSICN alongside each raw stress-peak locator to see the Gmsh SICN of that exact tetrahedron. Compare it with the mesh minimum only as local mesh-shape context; neither a high value nor separation from the minimum proves stress accuracy, a resolved hotspot, convergence, or strength. Read maximumPrincipalStressMPa and minimumPrincipalStressMPa as raw tensile/compressive principal extrema across sampled integration points, each with its own mesh locator. Their refinement trends and signed relative changes expose mesh sensitivity only; they are diagnostic and must never be compared with a von Mises allowable or reported as a pass without a criterion qualified for the selected failure mode. The public FEA tool measures the rank of all fixed global translations at the actual mapped mesh nodes and stops before CalculiX if they leave any of the six rigid-body translation/rotation modes unconstrained. A full rank of six is necessary to remove those rigid-body modes, but it does not establish physical support validity, elastic stability, or absence of local mechanisms. For a heat-set insert, use pullout and torque-out capacity only when the evidence matches the exact insert, host material, print profile, orientation, pocket and installation process. Ask for the worst-case demand on one insert; do not divide a group load evenly without a load-path model. For two or more fasteners under an in-plane load, establish every transfer-point coordinate, both force components, the point of application and any free moment. Use the elastic group method only after confirming a rigid attachment and identical in-plane fastener stiffness. Use each fastener’s own vector resultant for any member check. On a rectangular mounting face, call plasticity_check_fastener_group_layout with an explicit opposedFaceId to verify matching native perforated faces and exact plate thickness. This is geometry evidence only and does not establish a load path or capacity. plasticity_verify_fastener_group_plate_bearing compares each elastic per-fastener demand with a directly traceable, configuration-matched, already factored bearing allowable using exact measured thickness and hole diameter. It can additionally check a straight transverse net-tension section only when you provide the external tensile resultant separately, identify local X or Y as its axis, supply a distinct traceable factored tensile allowable, and confirm uniform membrane tension, centered through-thickness loading and a straight transverse failure path. Never substitute per-fastener demands for the external plate tension or infer that resultant from a sketch. Optionally use edgeShearOut with a separate traceable, factored shear allowable to check local two-plane tear-out for per-fastener vectors aligned to local X or Y; diagonal vectors and e/d below 1.5 are unsupported, while e/d below 2 remains conditional. Angled/staggered fracture paths, compression-side buckling, shared-ligament interaction, unsupported tear-out directions, bypass and complete-joint strength remain unchecked; even a within-allowable result is only a conditional local screen. If a physical multi-hole plate test is run, record the measured specimen, exact hole layout, print-process/profile, fixture, load axis, individual peak loads and observed failure modes with plasticity_record_fastener_group_test, then use plasticity_match_fastener_group_test to find only an exact configuration match. A test match is evidence only: it does not produce a design allowable or pass a different part or support setup. To add a test benchmark to the CAD-bound report, first confirm that the record's exact process and fixture/load path apply to the current part. Supply the selected immutable record ID, an independently evidenced dimensional equivalence tolerance, the exact process, safety factor, and explicit process/fixture confirmations; the server then checks every measured plate and hole dimension against the live B-rep. This comparison only checks factored external tensile demand against the lowest observed specimen peak. It is not a statistically reduced allowable, strength pass/fail, or proof for unobserved failure modes. Do not select a nearby test by appearance or silently assume fixture equivalence. The single-through-fastener plate method assumes one hole; never repeat it per hole and aggregate the results into a multi-hole plate or whole-joint pass. Calculate preliminary dimensions, then propose one logical CAD change in the chat. Execute only the accepted package or the current explicitly delegated task. Delegation ends when the task ends and is not restored after restart. Read actual native geometry and recalculate after manual edits or manufacturing changes. Workbench is optional. Unknown material properties cannot be converted into a pass by confidence language. For an end-to-end tongue-root scenario, use plasticity_calculate_tongue_root_strength_from_coupon_data or plasticity_verify_tongue_root_strength_from_coupon_data. They match the physical coupon record to the exact Workbench profile hash, printer, material, orientation, infill, nozzle temperature and measured slicer layer height, then copy only measured Young's/shear moduli and their evidence into the calculation. Supply separately sourced tensile/shear design allowables and explain their applicability; raw coupon strengths are never substituted for allowables. These tools keep material suitability unconfirmed and the result conditional. No-match or conflicting records return without a calculation. If no record exists, ask for the report and unresolved process details, and record it only after the user confirms the physical tests. The registry does not check test-standard compliance, derive statistical design values, or certify a part. Never substitute a generic filament label or slicer properties for coupon evidence. For axial rectangular or rectangular-cantilever scenarios with exact-process coupon data, use plasticity_calculate_rectangular_strength_from_coupon_data. It copies only measured Young's modulus and its evidence. Provide independent tensile design allowable evidence, plus an independent compressive allowable for a cantilever; state their applicability basis. Never map raw coupon tensile/compressive strengths into design allowables. These scenarios remain conditional with material suitability unconfirmed. Exact-process no-match or ambiguity returns without saving a calculation. Nominal section stress is not whole-part validation.

Re-read exact opposed planar faces of an integral enclosure wall, replace plate dimensions with measured native geometry, calculate the uniform-pressure plate scenario and persist a CAD-bound report. All four simple supports remain an explicit engineering assumption. Search accessible primary product/material sources before asking the user for known facts. Before requesting missing print-material data, call plasticity_plan_single_material_strength_tests for the selected method and exact one-material process. Use its measurement matrix to ask only for evidence needed by that route; explain why the measurements matter and never invent coupon values or allowables. A DCB task may contain an explicitly labeled generic-PLA literature geometry as a starting reference only; do not present it as a Creality property, normative specimen size, or sample-count requirement, and require checking the selected process, fixture and method. Keep the confirmed road/build axes and same-material layer-interface assumptions explicit. For Creality PLA literature references, read plasticity://strength/interlayer-literature-baseline. It contains separate CR-PLA and Hyper PLA records plus generic-PLA Z-tension/DCB context; do not merge product families or transfer values across SKUs. It is a screening reference only: do not register literature values as physical coupons or interface tests, bind them to the user's exact K1C process, treat them as design allowables, or infer a traction-separation curve from fracture energy. Manufacturer-mirror discrepancies and non-matching printer/process tests must remain visible; ask for exact-process physical tests before a calibrated cohesive result. When the user provides raw direct-tension machine CSV, call plasticity_import_interface_tensile_csv first with explicit specimen-ID/force headers, delimiter, decimal separator, force unit and sign, measured net area and observed failure location for every coupon. Review its hashed peak-force preview; the importer does not filter or correct machine data, register a test, or derive DCB/ENF/MMB curves. For directly loaded specimens supplied in another format, record every sample's measured peak force, net cross-section, observed failure location and SHA-256/source locator, then call plasticity_calculate_interface_specimen_strengths; report force/area as nominal specimen stress. Include only confirmed interface failures in its summary statistics; show bulk/fixture failures individually and do not pool them with interface failures. Keep all statistics descriptive only. Do not call these local interface tractions or design allowables, and do not use them as a cohesive law. When a raw DCB force/displacement CSV is available, use plasticity_import_dcb_mode_i_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting explicitly, then select the exact source record for every manually observed crack-growth point and enter that measured crack length; never let the importer filter traces, infer growth or select peak loads. It converts units only, preserves the source hash/record locators and calculates an exploratory MBT G_I(a) curve. Require the caller to establish machine-compliance-corrected load-point displacement and quasi-static linear-elastic behavior; attestations are not independent verification. Review selected rows, crack lengths, calculation and limitations against the physical log. If no CSV is available, plasticity_calculate_dcb_mode_i_energy accepts equivalent manually selected observations with traceable source hash/locators. This does not conform the test to ASTM D5528 (whose stated scope is unidirectional fiber composites), create a traction-separation law, or qualify CR-PLA. Never convert G_I into peak traction or strength without separate evidence. When raw ENF force/displacement CSV is available, use plasticity_import_enf_mode_ii_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting; group calibration rows into each measured run/crack length, manually select only the initial-linear records, and manually select the observed initiation/peak record for the fracture run. The tool fits compliance from the selected rows, then the existing ENF calculator fits C against a^3 and returns exploratory G_IIc; it never searches for the linear range, crack growth or peak. Review each compliance fit, source hash/record locator, force/displacement correction and fixture log. If pre-reduced compliances are supplied, use plasticity_calculate_enf_mode_ii_energy directly. Neither path determines ASTM D7905 validity for printed PLA, registers physical evidence, builds an R-curve or yields a cohesive law. Keep the confirmed in-plane shear direction exact; do not merge ENF runs with differing direction just because their layer normals match. When the requested failure mode is layer separation, distinguish CR-PLA's physical tensile/infill study and the two-sample CR-PLA-associated vertical layer-adhesion screen from the Hyper PLA flat tensile/flexural study: observed flexural delamination is qualitative failure-mode evidence, not a measured interface law. The baseline also records an upright generic-PLA tensile coupon on a Creality Ender 3 Pro and a custom generic-PLA interface coupon/calibrated cohesive parameter; neither identifies the filament as Creality nor establishes a same-process cohesive calibration. For only a nominal normal-strength screen across the layers, use the focused layer-interface-normal-tension scope; require measured failure at the interface and do not use that peak as a cohesive law. Use layer-interface-mode-i for a Mode-I fracture response, layer-interface-mode-ii for exploratory ENF Mode-II initiation energy using a same-fixture compliance calibration, and layer-interface-mixed-mode only when the requested analysis needs DCB/ENF/MMB cohesive evidence. The focused ENF calculator fits the experimentally supplied compliance against crack length cubed and uses the initiation peak; it is not an R-curve, does not verify the physical fixture or raw compliance fit, and does not claim ASTM D7905 validity for printed PLA. Layerwise static FEA assumes perfectly bonded interfaces and cannot answer a delamination question. After reviewing an exploratory DCB energy preview against the physical test log, persist it only through plasticity_record_dcb_mode_i_energy_test with explicit confirmation that the measurements came from real physical tests. Retrieve a full record with plasticity_read_dcb_mode_i_energy_test and require exact process/interface-normal/protocol matching through plasticity_match_dcb_mode_i_energy_test before comparing runs. For ENF Mode-II energy, record reviewed measurements only through plasticity_record_enf_mode_ii_energy_test with the same physical-test confirmation; retrieve all calibration/fracture inputs with plasticity_read_enf_mode_ii_energy_test and exact-match process/interface-normal/in-plane-shear-axis/protocol using plasticity_match_enf_mode_ii_energy_test. Both immutable energy registries are separate from peak-strength and traction-separation records and are not consumed by cohesive FEA. For an exploratory mixed-mode initiation partition, use plasticity_calculate_mmb_mode_i_ii_energy only when measured MMB force and geometry, same-process flexural and orthotropic moduli, and material-axis mapping are available; require lever weight to be measured negligible or counterbalanced. This Reeder-Crews beam-theory estimate does not establish ASTM D6671 validity for printed PLA and does not replace full mixed-mode traction-separation curves or provide a cohesive law. For raw MMB force CSV, use plasticity_import_mmb_mode_i_ii_energy_csv only with manually selected initiation records and a physically observed criterion; it does not search traces for onset/peak. Record a preview only after reviewing the source and confirming the data are from physical tests; matching caller confirmation with plasticity_record_mmb_mode_i_ii_energy_test persists and server-recomputes this separate energy estimate. Use plasticity_match_mmb_mode_i_ii_energy_test only for the exact process, interface normal, shear axis and protocol; read the full evidence with plasticity_read_mmb_mode_i_ii_energy_test. This registry is not cohesive input or an FEA source. When force is unknown, ask what object is supported, how it is mounted and used. Record source, units and uncertainty; do not infer exact scale from an unscaled image. Ask at most one next-step question package per response, then wait for the user's answer. Include only facts needed to choose the next safe step; defer material, manufacturing, tolerances and detailed dimensions until they affect that decision. Re-evaluate after every answer and ask a focused follow-up only when it changes the method, required evidence or next action. If the user does not know, move to one useful contextual clue such as the supported object, use, environment or mounting; do not repeat a list of unknowns or guess. On a new bracket task, first ask compactly what it supports, its load/use and how it is mounted; defer section, material/process and displacement questions until that first answer narrows the load path and method. Choose a supported member method and report its unchecked components explicitly. For a flat rectangular panel, establish net pressure and the real condition of all four edges before choosing the simply-supported plate method. Do not infer edge support from appearance. For a straight prismatic rectangular member in centred axial compression, use the Euler column method only when effective-length factor K, elastic limit, compressive allowable, and the actual restraint condition are supported by evidence. Use the weakest section axis, require a negative force value for compression, and reject Euler results when its predicted critical stress exceeds the supplied elastic limit; do not guess K or treat the method as an inelastic, eccentric-load, local-buckling, or whole-part check. For an integral rectangular enclosure wall, inspect two opposed planar faces on the same Solid with plasticity_inspect_integral_rectangular_plate, then use plasticity_verify_integral_plate_strength to replace dimensions from exact native B-rep. The inspector accepts only rectangular faces whose outlines match or inset by one wall thickness on each side. It measures geometry only: establish the real support, pressure, material and edge conditions separately; never assume a box wall is simply supported. For section analysis, identify the critical plane and explain the load path that makes it critical. Use an existing planar face when it matches; otherwise inspect an arbitrary plane through the Solid. For one fastener carrying in-plane plate load, establish the load direction and separate bearing, shear and tensile allowables before calculating. Never substitute compressive strength for bearing strength. For a fastener group on a rectangular planar face, inspect exact hole centers and boundaries, then check center-to-edge distance, hole-edge clearance, pitch, ligament and every applicable head, washer, nut or driver envelope against explicitly sourced or user-approved criteria. A measured layout without criteria is not a pass, and a layout pass is not a strength pass. When checking the fastener member, establish its grade, tensile stress area, effective shear area at the actual plane, one or two shear planes, total axial tension including applicable preload, and actual shear load. Explain why the selected combined-load interaction is applicable before accepting it. For a tapped hole, nut, or threaded insert under axial load, establish the designation, pitch, actual engagement, fully formed engaged thread count, and configuration-matched allowable loads for internal-thread stripping, external-thread stripping, and fastener tension. Do not infer any capacity from nominal M size. A procured nut or insert requires a specified assembly allowable or dedicated test rather than an unqualified nominal shear-area calculation. For solver-backed static FEA, first use plasticity_match_material_coupon_data when a physically tested material record may match the print process. Only an unambiguous exact match for printer, material, profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature and measured slicer layer height can supply Young's modulus; pass the selected record ID/process and the exact recorded modulus so the server binds and validates it. If match is ambiguous because compatible partial records have no single consolidated record, ask the user for the confirmed count of unique physical specimens and call plasticity_combine_material_coupon_data with those exact record IDs; it preserves measured values and evidence without averaging. If records conflict, stop and ask the user which physical measurement applies. For an orthotropic print model, do not use a uniaxial coupon as a full tensor. When an immutable exact-process record contains the complete measured tensor and nu12, pass orthotropicMaterial.couponRecordId and process, plus materialCoupon with the same record ID; the MCP checks E1, nu12, every remaining constant, each evidence object and the confirmed global print axes, and rechecks the record when reading the saved report. Otherwise pass E2/E3, nu13/nu23, G12/G13/G23 with separate exact evidence, global material axes 1 and 2, and a user-confirmed or sourced global build direction; material axis 3 must align with that build direction and represents the homogeneous layer-normal response. This does not represent individual layers or prove adhesion. Model one material and one print process per analyzed part; do not combine material datasets or infer a multi-material print. When the selected exact-process coupon record contains a qualified Tsai-Wu dataset, pass its ID as orthotropicMaterial.tsaiWuQualificationRecordId along with the identical orthotropicMaterial.process; the server loads measured strengths, biaxial interaction data and source evidence, then verifies the axes. Do not copy or reconstruct criterion values by hand. If the exact-process match remains ambiguous after consolidation, ask the user to resolve conflicting measurements; never choose a record silently. The legacy youngsModulusMPa and poissonRatio fields are E1 and nu12. Ask the user to resolve the print-axis orientation if it is unknown; never infer it from a photo or CAD face. The server checks positive-definite compliance and rejects a von Mises allowable or coupon binding for this orthotropic model. Otherwise attach youngsModulusEvidence whose value exactly matches the modulus; measured/sourced evidence needs a URL, SHA-256 and locator. poissonRatioEvidence is always required and must exactly match the ratio. A physical coupon record may store measured nu12 with its exact process. If using its Tsai-Wu record-binding path, the server verifies the nu12 value and evidence against that record; otherwise supply the input evidence directly from a traceable source. Never infer nu12 from a generic plastic label. An assumed value must include its reason and stay explicitly scenario-only after discussing the assumption with the user. Optionally supply a directly measured or sourced, process-applicable factored von Mises design allowable using factoredVonMisesAllowableMPa, exact matching factoredVonMisesAllowableEvidence with URL, SHA-256 and locator, and factoredVonMisesAllowableBasis. It must already include the design factors. Never turn a generic tensile strength or raw coupon peak into a design allowable. The report will compare every sampled raw mesh peak with it as a diagnostic screen only: above means a sampled peak exceeds that supplied allowable; below does not prove strength. This never changes strengthPass or print approval. Confirm the actual restraints with the user. Legacy supportFaceIds fix all three global translations on each selected planar face; use explicit supportConditions only when the user has specified which global x/y/z translations are zero on each face. Each condition acts on all nodes of that face, is not a frictionless-contact or rotational support, and may leave rigid-body modes or create a singular system; do not infer it from a photo or face normal. Supports and loaded faces must not share mesh nodes. Every selected planar load face needs an explicit uniform traction and/or resultant force with its point of application and free moment. Use plasticity_analyze_static_fem only for one Solid with 1–8 explicit support conditions and load vectors grounded in a selected planar face. Use named loadCases for physically distinct scenarios such as weight, operating force and handling load; do not combine mutually exclusive scenarios. Each case has a separate solver result, and comparison proceeds only when the generated meshes are byte-identical. Set meshRefinementSteps to 1–3 for two to four mesh levels when a trend across successively halved element sizes is useful; the report classifies the sampled direction of raw maximum von Mises stress and observed displacement as increasing, decreasing, unchanged, non-monotonic or insufficient-levels. For more than four levels, run separate analyses against the same current CAD revision and call plasticity_compare_static_fem_refinement_reports; it merges only reports with matching loads, supports, material evidence, native geometry and byte-identical overlapping mesh results. Check the returned freshness before relying on it. Treat all trend labels and relative changes as diagnostic evidence only. Never call a mesh trend a pass or proof of convergence. Each result includes the raw maximum C3D4 integration-point stress element and its mesh-element centroid in millimetres; this is a mesh-bound locator, not an averaged stress field, a resolved critical-region boundary, or proof of a physical hotspot. Locations may move between refinement levels. The legacy top-level load fields still represent one case. The tool returns total and per-support reaction forces and moments, global force/moment equilibrium residuals and a named displacement axis for each case; per-support values are diagnostic resultants over each support node set. Re-read it with plasticity_static_fem_report; any CAD revision change makes it stale.

For layerwise static FEA, keep the single-material exact process consistent from measurement through solver input. Read the immutable profile hash, infill pattern, wall loops, top/bottom shell layers and measured layer height from workbench_manufacturing_profiles, then slice the intended model/orientation with the selected Creality K1C profile. Record actual specimen settings if per-object overrides differ from profile defaults; an unchanged profile hash does not erase those differences. Call workbench_slicer_layer_path_orientations in batches of up to 32 and combine layer-path results for every deposited layer, including the final deposited layer, into layerPathEvidence with identical job ID, profile hash, source-artifact hash and G-code hash. Layerwise static orientation is supported only for a complete linear or planar circular-arc direction result for every layer in stacks of 2–33 layers; do not fill missing or curved-path directions by inference. Confirm pathFrameMapping.slicerXDirectionGlobal in CAD global coordinates and explicitly confirm that the same exact-process coupon's material axis 1 represents the dominant deposited-road direction in roadAxisMapping. Call plasticity_plan_cohesive_layer_planes with the same profile hash and layer height, current CAD anchor at the first interlayer plane, confirmed CAD build direction, total layer count, and the full layer-path evidence/mappings; pass its returned plan as layerPlanePlan to plasticity_analyze_static_fem. When the slicer supplies actual interface heights, call workbench_slicer_interface_heights for every interface in this complete stack, set each interfaceOffsetsMm value to its depositionLayerZMm minus firstDepositionLayerZMm, and attach the matching job/profile/source/G-code hashes as depositionPathEvidence; this preserves first-layer and adaptive heights instead of assuming nominal uniform spacing. The static analyzer requires the plan to match the one measured orthotropic process and applies the same measured orthotropic tensor to each layer with its confirmed G-code-mapped frame. It assumes perfectly bonded layers and cannot assess delamination. Do not use static FEA to assess delamination; use the separate same-material cohesive route only when matching physical interface tests are available. Never infer layer directions, material identity or interlayer strength from a photo, generic material label or nominal slicer preset.

For orthotropic FEA, optionally supply all nine directly traceable, already factored X/Y/Z tensile and compressive plus XY/XZ/YZ shear limits in orthotropicMaterial.factoredAllowables, with separate exact-value evidence and an applicability/design-factor basis. Never substitute generic datasheet strength or an unqualified coupon peak. The returned componentwise maximum-stress screen uses local material-axis stress extrema and is diagnostic only; it assumes one homogeneous orthotropic continuum, does not model layer interfaces, delamination or different-material joints, and omits multiaxial interaction. Matched interlayer-test data can inform Z-tension and XZ/YZ-shear allowables but does not turn this into a cohesive-interface analysis. Never report this screen as verified layer adhesion, part strength or print approval. For an optional 3D Tsai-Wu first-failure screen in orthotropic linear FEA, require one exact single-material printer/material/profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers and nozzle-temperature identity in orthotropicMaterial.process. Prefer binding the unique exact-process qualification by its orthotropicMaterial.tsaiWuQualificationRecordId; the server loads its nine directly measured, un-factored X/Y/Z tensile/compressive and XY/XZ/YZ shear failure strengths, three normalized XY/XZ/YZ normal-interaction coefficients and evidence, then verifies material axes. Each strength test must attest its material-frame axis and mode (for example, x tension is material-1 tension; xy shear is material-1-2 shear); each interaction and its source dependencies must attest the corresponding biaxial plane. Legacy records without these direction attestations remain readable but cannot qualify a Tsai-Wu FEA. A complete inline orthotropicMaterial.tsaiWuCriterion remains supported when needed. Interactions must be derived from traceable biaxial tests. Ask the user for these records if missing; never copy generic datasheet values, assume the conventional interaction coefficient or infer it from uniaxial coupons. The normalized interaction matrix must be positive definite. The result reports local integration-point failure indices and proportional load factors to index one only. A load factor is not a design safety factor, and neither a sub-unity index nor a large load factor means the part passed. The model still represents one homogeneous material, not individual roads or delamination.

When actual test-coupon G-code is available, preserve its profile/source/G-code hashes, selected per-layer road summaries and explicitly user-confirmed slicer-to-global axes in depositionPathEvidence. This is test provenance only, not an adhesion measurement or solver input; never reconstruct road direction from nominal slicer settings.

When physical adhesion between printed layers of one material is relevant, use plasticity_record_material_interface_test only for caller-provided measured test results with the same exact printer/material/profile process on both sides, interface normal, test mode, load direction, fixture/specimen protocol hash and observed failure location. Normal-tension load direction must align with the interface normal; interface-shear load direction must lie in its plane; mixed-mode must contain both components. Store full compliance-corrected pure-mode curves as scalar separation/traction data and MMB curves as separate normal/tangential separation and traction components, with source SHA-256 and locator. Call plasticity_analyze_material_interface_test_curve to summarize measured work and mode mixity. When the user has a CSV of already processed physical fracture data, use plasticity_import_interface_fracture_csv for a read-only per-specimen preview; explicitly map the method, columns, units and CSV formatting, and attest that the values are already compliance-corrected physical traction-separation data. Never convert raw machine force-displacement data with this importer. Verify each source hash/record locator, specimen, fixture, exact print process and observed failure plane before separately recording any curve with plasticity_record_material_interface_test. For a CZM_TURON candidate fit, provide Mode-I DCB, Mode-II ENF and at least two MMB tests at distinct measured energy fractions; all ENF and MMB tests must use the same in-plane shear axis within one degree because this route has a single tangential cohesive law. The MCP requires those protocol identifiers in each testMethod and reports the pure-mode peaks plus fit residuals. Review residuals against test uncertainty. Do not invent ETA_BK. K is not identified by that fit and must not be silently guessed. Code_Aster CZM_TURON uses one normal and one tangential cohesive response, sharing tangential strength and fracture energy across both in-plane tangent directions; it cannot represent direction-dependent shear adhesion. Matching shear axes prevents mixing directional datasets but does not prove that the bond is isotropic. These outputs summarize physical evidence only; they are not qualified cohesive-law parameters, design allowables or a part FEA. Do not claim bond integrity from the homogeneous orthotropic FEA screen. Use plasticity_analyze_cohesive_interface only for a single-material printed part, with an exact-process coupon for that one bulk material, traceable Poisson ratio and a matching immutable same-material layer test. Dissimilar-material printed bonds are rejected before meshing and are outside this calculation scope. Mode-I analysis requires a full normal-tension DCB traction-separation curve with failure at the layer interface. Its modeILaw may explicitly select CZM_EXP_REG or CZM_LIN_REG; omitted input preserves the CZM_EXP_REG default. Both parameterize their softening law from measured peak traction and integrated fracture energy and do not fit the full measured curve shape. Review this choice against the measured curve and record the reason; do not infer it from part geometry. Read the returned solver.result.v3Interpretation: for CZM_EXP_REG, V3 is a damage variable in [0,1]; for CZM_LIN_REG, V3=2 means the cohesive element is completely broken, so do not present it as the same normalized damage fraction. By default it uses isotropic bulk elasticity with pinned Code_Aster 15.2. If useOrthotropicBulkProperties is explicitly enabled, Mode-I uses Code_Aster 17.4 and applies one measured homogeneous orthotropic tensor; when explicit layerwise mapping is enabled, each bulk layer uses its G-code-mapped frame while all layers share that same tensor. The mixed-mode Turon route additionally requires same-process-pair DCB, ENF and at least two MMB tests at distinct measured energy fractions, traceable K, and an explicit global displacement with both opening-normal and in-plane tangential components greater than one degree. The displacement's shear axis must match the common measured ENF/MMB shear axis within one degree; reject other directions because this solver law has one tangential response. All ENF and MMB tests must use the same in-plane shear axis within one degree. Supply 1..32 ordered parallel split planes for a same-material layer stack; their normals may be tilted in global coordinates, and the same measured layer law is repeated at every plane. This equivalent-interface assumption does not resolve individual roads or within-layer raster variation. Either route may accept useOrthotropicBulkProperties when the exact-process coupon contains measured E2/E3, nu13/nu23, G12/G13/G23 evidence and a confirmed or traceable global print frame. Check the exact-process coupon match is unambiguous and do not infer axes from a photo or CAD face. Confirm the tested material's side relative to the chosen split-plane normal; do not infer this from the body or face normals. The measured test direction must match the normal/shear mode; the support face lies below the first split plane and the loaded face above the last plane along the shared normal. The server derives peak traction, integrated fracture energy and displacement endpoint from exact tests, exports the selected current Solid, creates a conforming multi-region mesh with a cohesive element set at every requested plane, and aborts if the CAD revision changes. Review the mesh-resolution screen and perform refinement/sensitivity work as needed. Turon approximates measured curves using peak/area parameters and its K still needs sensitivity analysis. The bulk tensor remains one measured single-material continuum whose local frame may vary by mapped layer; repeated cohesive planes use the same measured, direction-independent interface law and do not resolve individual deposited roads, within-layer raster mixtures or direction-dependent adhesion. Use results only as raw solver responses; never report it as a part-strength verdict, qualified layer-adhesion value, print approval or design allowable. Before asking for a physical coupon record, obtain the selected immutable profile hash plus material.nozzleTemperatureC, slicer.layerHeightMm, slicer.nominalInfillPercent, slicer.sparseInfillPattern, slicer.wallLoops, slicer.topShellLayers, and slicer.bottomShellLayers from workbench_manufacturing_profiles when those settings are available. Carry the exact process values into the record; do not infer missing infill or substitute a generic profile. These are resolved profile defaults; sparse fill or shell settings may be overridden per object and must not be assumed when a modifier was used. If the coupon was printed with a different setting, first register/select the matching immutable process profile and hash. When layer positions should follow the actual print, record the measured profile hash and layer height in the physical interface-test and material-coupon process; layer height is part of exact process identity. After slicing, if the job reports a complete deposition-height schedule, call workbench_slicer_interface_heights for every selected interface index. Derive each offset as that interface’s depositionLayerZMm minus firstDepositionLayerZMm and pass the ordered values as interfaceOffsetsMm; this preserves first-layer and adaptive heights. Also pass the selected interface response plus its job/profile/source/G-code hashes, layer count and coordinate frame as depositionPathEvidence so the cohesive report retains the exact per-layer road-orientation observations. This evidence preserves the measured toolpath but does not qualify material properties. Use workbench_slicer_layer_path_orientations in batches of up to 32 to retrieve every layer direction (up to 33 total layers) when the user wants layerwise solver orientation. Confirm how slicer X maps into the CAD global frame and that the exact-process coupon axis 1 represents the dominant deposited-road direction; pass those confirmations as pathFrameMapping and roadAxisMapping. Combine complete responses, including the final deposited layer, as layerPathEvidence with shared job/profile/source/G-code hashes. Every layer must have complete linear or planar circular-arc coverage; do not treat unsupported arc/spline moves as complete. With explicit roadAxisMapping and useOrthotropicBulkProperties, Mode-I and Turon use one measured tensor with a separate local frame per layer. Without roadAxisMapping, frames remain candidates and do not affect solver response. This does not model multiple materials, layer-varying properties, within-layer raster mixtures or directional interface adhesion. Returned coordinates are in slicer build coordinates: map only relative offsets onto the confirmed CAD print axis, never copy absolute slicer Z into CAD. Otherwise the planner uses the nominal profile height. Call plasticity_plan_cohesive_layer_planes with that same profile hash and layer height, a point on the first interlayer plane after object placement, confirmed global build direction, total layer count and explicit interface indices. Copy both its plan and planes into layerPlanePlan and splitPlanes; analysis rejects legacy test records without a recorded layer height and any height that differs from the measured process profile. It also checks profile identity, plane coordinates, and (for orthotropic bulk) alignment with the coupon’s confirmed build direction. For up to 32 interfaces the planner requires the complete stack; for larger stacks it marks selected planes as incomplete, and omitted interfaces remain unanalyzed. Do not present selected-plane analysis as full-stack delamination resistance. If no validated placement/anchor is available, ask instead of inferring layer positions from an image or display mesh. Read maximumVonMisesElementSICN alongside each raw stress-peak locator to see the Gmsh SICN of that exact tetrahedron. Compare it with the mesh minimum only as local mesh-shape context; neither a high value nor separation from the minimum proves stress accuracy, a resolved hotspot, convergence, or strength. Read maximumPrincipalStressMPa and minimumPrincipalStressMPa as raw tensile/compressive principal extrema across sampled integration points, each with its own mesh locator. Their refinement trends and signed relative changes expose mesh sensitivity only; they are diagnostic and must never be compared with a von Mises allowable or reported as a pass without a criterion qualified for the selected failure mode. The public FEA tool measures the rank of all fixed global translations at the actual mapped mesh nodes and stops before CalculiX if they leave any of the six rigid-body translation/rotation modes unconstrained. A full rank of six is necessary to remove those rigid-body modes, but it does not establish physical support validity, elastic stability, or absence of local mechanisms. For a heat-set insert, use pullout and torque-out capacity only when the evidence matches the exact insert, host material, print profile, orientation, pocket and installation process. Ask for the worst-case demand on one insert; do not divide a group load evenly without a load-path model. For two or more fasteners under an in-plane load, establish every transfer-point coordinate, both force components, the point of application and any free moment. Use the elastic group method only after confirming a rigid attachment and identical in-plane fastener stiffness. Use each fastener’s own vector resultant for any member check. On a rectangular mounting face, call plasticity_check_fastener_group_layout with an explicit opposedFaceId to verify matching native perforated faces and exact plate thickness. This is geometry evidence only and does not establish a load path or capacity. plasticity_verify_fastener_group_plate_bearing compares each elastic per-fastener demand with a directly traceable, configuration-matched, already factored bearing allowable using exact measured thickness and hole diameter. It can additionally check a straight transverse net-tension section only when you provide the external tensile resultant separately, identify local X or Y as its axis, supply a distinct traceable factored tensile allowable, and confirm uniform membrane tension, centered through-thickness loading and a straight transverse failure path. Never substitute per-fastener demands for the external plate tension or infer that resultant from a sketch. Optionally use edgeShearOut with a separate traceable, factored shear allowable to check local two-plane tear-out for per-fastener vectors aligned to local X or Y; diagonal vectors and e/d below 1.5 are unsupported, while e/d below 2 remains conditional. Angled/staggered fracture paths, compression-side buckling, shared-ligament interaction, unsupported tear-out directions, bypass and complete-joint strength remain unchecked; even a within-allowable result is only a conditional local screen. If a physical multi-hole plate test is run, record the measured specimen, exact hole layout, print-process/profile, fixture, load axis, individual peak loads and observed failure modes with plasticity_record_fastener_group_test, then use plasticity_match_fastener_group_test to find only an exact configuration match. A test match is evidence only: it does not produce a design allowable or pass a different part or support setup. To add a test benchmark to the CAD-bound report, first confirm that the record's exact process and fixture/load path apply to the current part. Supply the selected immutable record ID, an independently evidenced dimensional equivalence tolerance, the exact process, safety factor, and explicit process/fixture confirmations; the server then checks every measured plate and hole dimension against the live B-rep. This comparison only checks factored external tensile demand against the lowest observed specimen peak. It is not a statistically reduced allowable, strength pass/fail, or proof for unobserved failure modes. Do not select a nearby test by appearance or silently assume fixture equivalence. The single-through-fastener plate method assumes one hole; never repeat it per hole and aggregate the results into a multi-hole plate or whole-joint pass. Calculate preliminary dimensions, then propose one logical CAD change in the chat. Execute only the accepted package or the current explicitly delegated task. Delegation ends when the task ends and is not restored after restart. Read actual native geometry and recalculate after manual edits or manufacturing changes. Workbench is optional. Unknown material properties cannot be converted into a pass by confidence language. For an end-to-end tongue-root scenario, use plasticity_calculate_tongue_root_strength_from_coupon_data or plasticity_verify_tongue_root_strength_from_coupon_data. They match the physical coupon record to the exact Workbench profile hash, printer, material, orientation, infill, nozzle temperature and measured slicer layer height, then copy only measured Young's/shear moduli and their evidence into the calculation. Supply separately sourced tensile/shear design allowables and explain their applicability; raw coupon strengths are never substituted for allowables. These tools keep material suitability unconfirmed and the result conditional. No-match or conflicting records return without a calculation. If no record exists, ask for the report and unresolved process details, and record it only after the user confirms the physical tests. The registry does not check test-standard compliance, derive statistical design values, or certify a part. Never substitute a generic filament label or slicer properties for coupon evidence. For axial rectangular or rectangular-cantilever scenarios with exact-process coupon data, use plasticity_calculate_rectangular_strength_from_coupon_data. It copies only measured Young's modulus and its evidence. Provide independent tensile design allowable evidence, plus an independent compressive allowable for a cantilever; state their applicability basis. Never map raw coupon tensile/compressive strengths into design allowables. These scenarios remain conditional with material suitability unconfirmed. Exact-process no-match or ambiguity returns without saving a calculation. Nominal section stress is not whole-part validation.

Re-read an exact native planar face or arbitrary plane through a Solid, replace section geometry with measured evidence, calculate and persist a CAD-bound report. It does not mutate persistent CAD geometry. Search accessible primary product/material sources before asking the user for known facts. Before requesting missing print-material data, call plasticity_plan_single_material_strength_tests for the selected method and exact one-material process. Use its measurement matrix to ask only for evidence needed by that route; explain why the measurements matter and never invent coupon values or allowables. A DCB task may contain an explicitly labeled generic-PLA literature geometry as a starting reference only; do not present it as a Creality property, normative specimen size, or sample-count requirement, and require checking the selected process, fixture and method. Keep the confirmed road/build axes and same-material layer-interface assumptions explicit. For Creality PLA literature references, read plasticity://strength/interlayer-literature-baseline. It contains separate CR-PLA and Hyper PLA records plus generic-PLA Z-tension/DCB context; do not merge product families or transfer values across SKUs. It is a screening reference only: do not register literature values as physical coupons or interface tests, bind them to the user's exact K1C process, treat them as design allowables, or infer a traction-separation curve from fracture energy. Manufacturer-mirror discrepancies and non-matching printer/process tests must remain visible; ask for exact-process physical tests before a calibrated cohesive result. When the user provides raw direct-tension machine CSV, call plasticity_import_interface_tensile_csv first with explicit specimen-ID/force headers, delimiter, decimal separator, force unit and sign, measured net area and observed failure location for every coupon. Review its hashed peak-force preview; the importer does not filter or correct machine data, register a test, or derive DCB/ENF/MMB curves. For directly loaded specimens supplied in another format, record every sample's measured peak force, net cross-section, observed failure location and SHA-256/source locator, then call plasticity_calculate_interface_specimen_strengths; report force/area as nominal specimen stress. Include only confirmed interface failures in its summary statistics; show bulk/fixture failures individually and do not pool them with interface failures. Keep all statistics descriptive only. Do not call these local interface tractions or design allowables, and do not use them as a cohesive law. When a raw DCB force/displacement CSV is available, use plasticity_import_dcb_mode_i_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting explicitly, then select the exact source record for every manually observed crack-growth point and enter that measured crack length; never let the importer filter traces, infer growth or select peak loads. It converts units only, preserves the source hash/record locators and calculates an exploratory MBT G_I(a) curve. Require the caller to establish machine-compliance-corrected load-point displacement and quasi-static linear-elastic behavior; attestations are not independent verification. Review selected rows, crack lengths, calculation and limitations against the physical log. If no CSV is available, plasticity_calculate_dcb_mode_i_energy accepts equivalent manually selected observations with traceable source hash/locators. This does not conform the test to ASTM D5528 (whose stated scope is unidirectional fiber composites), create a traction-separation law, or qualify CR-PLA. Never convert G_I into peak traction or strength without separate evidence. When raw ENF force/displacement CSV is available, use plasticity_import_enf_mode_ii_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting; group calibration rows into each measured run/crack length, manually select only the initial-linear records, and manually select the observed initiation/peak record for the fracture run. The tool fits compliance from the selected rows, then the existing ENF calculator fits C against a^3 and returns exploratory G_IIc; it never searches for the linear range, crack growth or peak. Review each compliance fit, source hash/record locator, force/displacement correction and fixture log. If pre-reduced compliances are supplied, use plasticity_calculate_enf_mode_ii_energy directly. Neither path determines ASTM D7905 validity for printed PLA, registers physical evidence, builds an R-curve or yields a cohesive law. Keep the confirmed in-plane shear direction exact; do not merge ENF runs with differing direction just because their layer normals match. When the requested failure mode is layer separation, distinguish CR-PLA's physical tensile/infill study and the two-sample CR-PLA-associated vertical layer-adhesion screen from the Hyper PLA flat tensile/flexural study: observed flexural delamination is qualitative failure-mode evidence, not a measured interface law. The baseline also records an upright generic-PLA tensile coupon on a Creality Ender 3 Pro and a custom generic-PLA interface coupon/calibrated cohesive parameter; neither identifies the filament as Creality nor establishes a same-process cohesive calibration. For only a nominal normal-strength screen across the layers, use the focused layer-interface-normal-tension scope; require measured failure at the interface and do not use that peak as a cohesive law. Use layer-interface-mode-i for a Mode-I fracture response, layer-interface-mode-ii for exploratory ENF Mode-II initiation energy using a same-fixture compliance calibration, and layer-interface-mixed-mode only when the requested analysis needs DCB/ENF/MMB cohesive evidence. The focused ENF calculator fits the experimentally supplied compliance against crack length cubed and uses the initiation peak; it is not an R-curve, does not verify the physical fixture or raw compliance fit, and does not claim ASTM D7905 validity for printed PLA. Layerwise static FEA assumes perfectly bonded interfaces and cannot answer a delamination question. After reviewing an exploratory DCB energy preview against the physical test log, persist it only through plasticity_record_dcb_mode_i_energy_test with explicit confirmation that the measurements came from real physical tests. Retrieve a full record with plasticity_read_dcb_mode_i_energy_test and require exact process/interface-normal/protocol matching through plasticity_match_dcb_mode_i_energy_test before comparing runs. For ENF Mode-II energy, record reviewed measurements only through plasticity_record_enf_mode_ii_energy_test with the same physical-test confirmation; retrieve all calibration/fracture inputs with plasticity_read_enf_mode_ii_energy_test and exact-match process/interface-normal/in-plane-shear-axis/protocol using plasticity_match_enf_mode_ii_energy_test. Both immutable energy registries are separate from peak-strength and traction-separation records and are not consumed by cohesive FEA. For an exploratory mixed-mode initiation partition, use plasticity_calculate_mmb_mode_i_ii_energy only when measured MMB force and geometry, same-process flexural and orthotropic moduli, and material-axis mapping are available; require lever weight to be measured negligible or counterbalanced. This Reeder-Crews beam-theory estimate does not establish ASTM D6671 validity for printed PLA and does not replace full mixed-mode traction-separation curves or provide a cohesive law. For raw MMB force CSV, use plasticity_import_mmb_mode_i_ii_energy_csv only with manually selected initiation records and a physically observed criterion; it does not search traces for onset/peak. Record a preview only after reviewing the source and confirming the data are from physical tests; matching caller confirmation with plasticity_record_mmb_mode_i_ii_energy_test persists and server-recomputes this separate energy estimate. Use plasticity_match_mmb_mode_i_ii_energy_test only for the exact process, interface normal, shear axis and protocol; read the full evidence with plasticity_read_mmb_mode_i_ii_energy_test. This registry is not cohesive input or an FEA source. When force is unknown, ask what object is supported, how it is mounted and used. Record source, units and uncertainty; do not infer exact scale from an unscaled image. Ask at most one next-step question package per response, then wait for the user's answer. Include only facts needed to choose the next safe step; defer material, manufacturing, tolerances and detailed dimensions until they affect that decision. Re-evaluate after every answer and ask a focused follow-up only when it changes the method, required evidence or next action. If the user does not know, move to one useful contextual clue such as the supported object, use, environment or mounting; do not repeat a list of unknowns or guess. On a new bracket task, first ask compactly what it supports, its load/use and how it is mounted; defer section, material/process and displacement questions until that first answer narrows the load path and method. Choose a supported member method and report its unchecked components explicitly. For a flat rectangular panel, establish net pressure and the real condition of all four edges before choosing the simply-supported plate method. Do not infer edge support from appearance. For a straight prismatic rectangular member in centred axial compression, use the Euler column method only when effective-length factor K, elastic limit, compressive allowable, and the actual restraint condition are supported by evidence. Use the weakest section axis, require a negative force value for compression, and reject Euler results when its predicted critical stress exceeds the supplied elastic limit; do not guess K or treat the method as an inelastic, eccentric-load, local-buckling, or whole-part check. For an integral rectangular enclosure wall, inspect two opposed planar faces on the same Solid with plasticity_inspect_integral_rectangular_plate, then use plasticity_verify_integral_plate_strength to replace dimensions from exact native B-rep. The inspector accepts only rectangular faces whose outlines match or inset by one wall thickness on each side. It measures geometry only: establish the real support, pressure, material and edge conditions separately; never assume a box wall is simply supported. For section analysis, identify the critical plane and explain the load path that makes it critical. Use an existing planar face when it matches; otherwise inspect an arbitrary plane through the Solid. For one fastener carrying in-plane plate load, establish the load direction and separate bearing, shear and tensile allowables before calculating. Never substitute compressive strength for bearing strength. For a fastener group on a rectangular planar face, inspect exact hole centers and boundaries, then check center-to-edge distance, hole-edge clearance, pitch, ligament and every applicable head, washer, nut or driver envelope against explicitly sourced or user-approved criteria. A measured layout without criteria is not a pass, and a layout pass is not a strength pass. When checking the fastener member, establish its grade, tensile stress area, effective shear area at the actual plane, one or two shear planes, total axial tension including applicable preload, and actual shear load. Explain why the selected combined-load interaction is applicable before accepting it. For a tapped hole, nut, or threaded insert under axial load, establish the designation, pitch, actual engagement, fully formed engaged thread count, and configuration-matched allowable loads for internal-thread stripping, external-thread stripping, and fastener tension. Do not infer any capacity from nominal M size. A procured nut or insert requires a specified assembly allowable or dedicated test rather than an unqualified nominal shear-area calculation. For solver-backed static FEA, first use plasticity_match_material_coupon_data when a physically tested material record may match the print process. Only an unambiguous exact match for printer, material, profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature and measured slicer layer height can supply Young's modulus; pass the selected record ID/process and the exact recorded modulus so the server binds and validates it. If match is ambiguous because compatible partial records have no single consolidated record, ask the user for the confirmed count of unique physical specimens and call plasticity_combine_material_coupon_data with those exact record IDs; it preserves measured values and evidence without averaging. If records conflict, stop and ask the user which physical measurement applies. For an orthotropic print model, do not use a uniaxial coupon as a full tensor. When an immutable exact-process record contains the complete measured tensor and nu12, pass orthotropicMaterial.couponRecordId and process, plus materialCoupon with the same record ID; the MCP checks E1, nu12, every remaining constant, each evidence object and the confirmed global print axes, and rechecks the record when reading the saved report. Otherwise pass E2/E3, nu13/nu23, G12/G13/G23 with separate exact evidence, global material axes 1 and 2, and a user-confirmed or sourced global build direction; material axis 3 must align with that build direction and represents the homogeneous layer-normal response. This does not represent individual layers or prove adhesion. Model one material and one print process per analyzed part; do not combine material datasets or infer a multi-material print. When the selected exact-process coupon record contains a qualified Tsai-Wu dataset, pass its ID as orthotropicMaterial.tsaiWuQualificationRecordId along with the identical orthotropicMaterial.process; the server loads measured strengths, biaxial interaction data and source evidence, then verifies the axes. Do not copy or reconstruct criterion values by hand. If the exact-process match remains ambiguous after consolidation, ask the user to resolve conflicting measurements; never choose a record silently. The legacy youngsModulusMPa and poissonRatio fields are E1 and nu12. Ask the user to resolve the print-axis orientation if it is unknown; never infer it from a photo or CAD face. The server checks positive-definite compliance and rejects a von Mises allowable or coupon binding for this orthotropic model. Otherwise attach youngsModulusEvidence whose value exactly matches the modulus; measured/sourced evidence needs a URL, SHA-256 and locator. poissonRatioEvidence is always required and must exactly match the ratio. A physical coupon record may store measured nu12 with its exact process. If using its Tsai-Wu record-binding path, the server verifies the nu12 value and evidence against that record; otherwise supply the input evidence directly from a traceable source. Never infer nu12 from a generic plastic label. An assumed value must include its reason and stay explicitly scenario-only after discussing the assumption with the user. Optionally supply a directly measured or sourced, process-applicable factored von Mises design allowable using factoredVonMisesAllowableMPa, exact matching factoredVonMisesAllowableEvidence with URL, SHA-256 and locator, and factoredVonMisesAllowableBasis. It must already include the design factors. Never turn a generic tensile strength or raw coupon peak into a design allowable. The report will compare every sampled raw mesh peak with it as a diagnostic screen only: above means a sampled peak exceeds that supplied allowable; below does not prove strength. This never changes strengthPass or print approval. Confirm the actual restraints with the user. Legacy supportFaceIds fix all three global translations on each selected planar face; use explicit supportConditions only when the user has specified which global x/y/z translations are zero on each face. Each condition acts on all nodes of that face, is not a frictionless-contact or rotational support, and may leave rigid-body modes or create a singular system; do not infer it from a photo or face normal. Supports and loaded faces must not share mesh nodes. Every selected planar load face needs an explicit uniform traction and/or resultant force with its point of application and free moment. Use plasticity_analyze_static_fem only for one Solid with 1–8 explicit support conditions and load vectors grounded in a selected planar face. Use named loadCases for physically distinct scenarios such as weight, operating force and handling load; do not combine mutually exclusive scenarios. Each case has a separate solver result, and comparison proceeds only when the generated meshes are byte-identical. Set meshRefinementSteps to 1–3 for two to four mesh levels when a trend across successively halved element sizes is useful; the report classifies the sampled direction of raw maximum von Mises stress and observed displacement as increasing, decreasing, unchanged, non-monotonic or insufficient-levels. For more than four levels, run separate analyses against the same current CAD revision and call plasticity_compare_static_fem_refinement_reports; it merges only reports with matching loads, supports, material evidence, native geometry and byte-identical overlapping mesh results. Check the returned freshness before relying on it. Treat all trend labels and relative changes as diagnostic evidence only. Never call a mesh trend a pass or proof of convergence. Each result includes the raw maximum C3D4 integration-point stress element and its mesh-element centroid in millimetres; this is a mesh-bound locator, not an averaged stress field, a resolved critical-region boundary, or proof of a physical hotspot. Locations may move between refinement levels. The legacy top-level load fields still represent one case. The tool returns total and per-support reaction forces and moments, global force/moment equilibrium residuals and a named displacement axis for each case; per-support values are diagnostic resultants over each support node set. Re-read it with plasticity_static_fem_report; any CAD revision change makes it stale.

For layerwise static FEA, keep the single-material exact process consistent from measurement through solver input. Read the immutable profile hash, infill pattern, wall loops, top/bottom shell layers and measured layer height from workbench_manufacturing_profiles, then slice the intended model/orientation with the selected Creality K1C profile. Record actual specimen settings if per-object overrides differ from profile defaults; an unchanged profile hash does not erase those differences. Call workbench_slicer_layer_path_orientations in batches of up to 32 and combine layer-path results for every deposited layer, including the final deposited layer, into layerPathEvidence with identical job ID, profile hash, source-artifact hash and G-code hash. Layerwise static orientation is supported only for a complete linear or planar circular-arc direction result for every layer in stacks of 2–33 layers; do not fill missing or curved-path directions by inference. Confirm pathFrameMapping.slicerXDirectionGlobal in CAD global coordinates and explicitly confirm that the same exact-process coupon's material axis 1 represents the dominant deposited-road direction in roadAxisMapping. Call plasticity_plan_cohesive_layer_planes with the same profile hash and layer height, current CAD anchor at the first interlayer plane, confirmed CAD build direction, total layer count, and the full layer-path evidence/mappings; pass its returned plan as layerPlanePlan to plasticity_analyze_static_fem. When the slicer supplies actual interface heights, call workbench_slicer_interface_heights for every interface in this complete stack, set each interfaceOffsetsMm value to its depositionLayerZMm minus firstDepositionLayerZMm, and attach the matching job/profile/source/G-code hashes as depositionPathEvidence; this preserves first-layer and adaptive heights instead of assuming nominal uniform spacing. The static analyzer requires the plan to match the one measured orthotropic process and applies the same measured orthotropic tensor to each layer with its confirmed G-code-mapped frame. It assumes perfectly bonded layers and cannot assess delamination. Do not use static FEA to assess delamination; use the separate same-material cohesive route only when matching physical interface tests are available. Never infer layer directions, material identity or interlayer strength from a photo, generic material label or nominal slicer preset.

For orthotropic FEA, optionally supply all nine directly traceable, already factored X/Y/Z tensile and compressive plus XY/XZ/YZ shear limits in orthotropicMaterial.factoredAllowables, with separate exact-value evidence and an applicability/design-factor basis. Never substitute generic datasheet strength or an unqualified coupon peak. The returned componentwise maximum-stress screen uses local material-axis stress extrema and is diagnostic only; it assumes one homogeneous orthotropic continuum, does not model layer interfaces, delamination or different-material joints, and omits multiaxial interaction. Matched interlayer-test data can inform Z-tension and XZ/YZ-shear allowables but does not turn this into a cohesive-interface analysis. Never report this screen as verified layer adhesion, part strength or print approval. For an optional 3D Tsai-Wu first-failure screen in orthotropic linear FEA, require one exact single-material printer/material/profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers and nozzle-temperature identity in orthotropicMaterial.process. Prefer binding the unique exact-process qualification by its orthotropicMaterial.tsaiWuQualificationRecordId; the server loads its nine directly measured, un-factored X/Y/Z tensile/compressive and XY/XZ/YZ shear failure strengths, three normalized XY/XZ/YZ normal-interaction coefficients and evidence, then verifies material axes. Each strength test must attest its material-frame axis and mode (for example, x tension is material-1 tension; xy shear is material-1-2 shear); each interaction and its source dependencies must attest the corresponding biaxial plane. Legacy records without these direction attestations remain readable but cannot qualify a Tsai-Wu FEA. A complete inline orthotropicMaterial.tsaiWuCriterion remains supported when needed. Interactions must be derived from traceable biaxial tests. Ask the user for these records if missing; never copy generic datasheet values, assume the conventional interaction coefficient or infer it from uniaxial coupons. The normalized interaction matrix must be positive definite. The result reports local integration-point failure indices and proportional load factors to index one only. A load factor is not a design safety factor, and neither a sub-unity index nor a large load factor means the part passed. The model still represents one homogeneous material, not individual roads or delamination.

When actual test-coupon G-code is available, preserve its profile/source/G-code hashes, selected per-layer road summaries and explicitly user-confirmed slicer-to-global axes in depositionPathEvidence. This is test provenance only, not an adhesion measurement or solver input; never reconstruct road direction from nominal slicer settings.

When physical adhesion between printed layers of one material is relevant, use plasticity_record_material_interface_test only for caller-provided measured test results with the same exact printer/material/profile process on both sides, interface normal, test mode, load direction, fixture/specimen protocol hash and observed failure location. Normal-tension load direction must align with the interface normal; interface-shear load direction must lie in its plane; mixed-mode must contain both components. Store full compliance-corrected pure-mode curves as scalar separation/traction data and MMB curves as separate normal/tangential separation and traction components, with source SHA-256 and locator. Call plasticity_analyze_material_interface_test_curve to summarize measured work and mode mixity. When the user has a CSV of already processed physical fracture data, use plasticity_import_interface_fracture_csv for a read-only per-specimen preview; explicitly map the method, columns, units and CSV formatting, and attest that the values are already compliance-corrected physical traction-separation data. Never convert raw machine force-displacement data with this importer. Verify each source hash/record locator, specimen, fixture, exact print process and observed failure plane before separately recording any curve with plasticity_record_material_interface_test. For a CZM_TURON candidate fit, provide Mode-I DCB, Mode-II ENF and at least two MMB tests at distinct measured energy fractions; all ENF and MMB tests must use the same in-plane shear axis within one degree because this route has a single tangential cohesive law. The MCP requires those protocol identifiers in each testMethod and reports the pure-mode peaks plus fit residuals. Review residuals against test uncertainty. Do not invent ETA_BK. K is not identified by that fit and must not be silently guessed. Code_Aster CZM_TURON uses one normal and one tangential cohesive response, sharing tangential strength and fracture energy across both in-plane tangent directions; it cannot represent direction-dependent shear adhesion. Matching shear axes prevents mixing directional datasets but does not prove that the bond is isotropic. These outputs summarize physical evidence only; they are not qualified cohesive-law parameters, design allowables or a part FEA. Do not claim bond integrity from the homogeneous orthotropic FEA screen. Use plasticity_analyze_cohesive_interface only for a single-material printed part, with an exact-process coupon for that one bulk material, traceable Poisson ratio and a matching immutable same-material layer test. Dissimilar-material printed bonds are rejected before meshing and are outside this calculation scope. Mode-I analysis requires a full normal-tension DCB traction-separation curve with failure at the layer interface. Its modeILaw may explicitly select CZM_EXP_REG or CZM_LIN_REG; omitted input preserves the CZM_EXP_REG default. Both parameterize their softening law from measured peak traction and integrated fracture energy and do not fit the full measured curve shape. Review this choice against the measured curve and record the reason; do not infer it from part geometry. Read the returned solver.result.v3Interpretation: for CZM_EXP_REG, V3 is a damage variable in [0,1]; for CZM_LIN_REG, V3=2 means the cohesive element is completely broken, so do not present it as the same normalized damage fraction. By default it uses isotropic bulk elasticity with pinned Code_Aster 15.2. If useOrthotropicBulkProperties is explicitly enabled, Mode-I uses Code_Aster 17.4 and applies one measured homogeneous orthotropic tensor; when explicit layerwise mapping is enabled, each bulk layer uses its G-code-mapped frame while all layers share that same tensor. The mixed-mode Turon route additionally requires same-process-pair DCB, ENF and at least two MMB tests at distinct measured energy fractions, traceable K, and an explicit global displacement with both opening-normal and in-plane tangential components greater than one degree. The displacement's shear axis must match the common measured ENF/MMB shear axis within one degree; reject other directions because this solver law has one tangential response. All ENF and MMB tests must use the same in-plane shear axis within one degree. Supply 1..32 ordered parallel split planes for a same-material layer stack; their normals may be tilted in global coordinates, and the same measured layer law is repeated at every plane. This equivalent-interface assumption does not resolve individual roads or within-layer raster variation. Either route may accept useOrthotropicBulkProperties when the exact-process coupon contains measured E2/E3, nu13/nu23, G12/G13/G23 evidence and a confirmed or traceable global print frame. Check the exact-process coupon match is unambiguous and do not infer axes from a photo or CAD face. Confirm the tested material's side relative to the chosen split-plane normal; do not infer this from the body or face normals. The measured test direction must match the normal/shear mode; the support face lies below the first split plane and the loaded face above the last plane along the shared normal. The server derives peak traction, integrated fracture energy and displacement endpoint from exact tests, exports the selected current Solid, creates a conforming multi-region mesh with a cohesive element set at every requested plane, and aborts if the CAD revision changes. Review the mesh-resolution screen and perform refinement/sensitivity work as needed. Turon approximates measured curves using peak/area parameters and its K still needs sensitivity analysis. The bulk tensor remains one measured single-material continuum whose local frame may vary by mapped layer; repeated cohesive planes use the same measured, direction-independent interface law and do not resolve individual deposited roads, within-layer raster mixtures or direction-dependent adhesion. Use results only as raw solver responses; never report it as a part-strength verdict, qualified layer-adhesion value, print approval or design allowable. Before asking for a physical coupon record, obtain the selected immutable profile hash plus material.nozzleTemperatureC, slicer.layerHeightMm, slicer.nominalInfillPercent, slicer.sparseInfillPattern, slicer.wallLoops, slicer.topShellLayers, and slicer.bottomShellLayers from workbench_manufacturing_profiles when those settings are available. Carry the exact process values into the record; do not infer missing infill or substitute a generic profile. These are resolved profile defaults; sparse fill or shell settings may be overridden per object and must not be assumed when a modifier was used. If the coupon was printed with a different setting, first register/select the matching immutable process profile and hash. When layer positions should follow the actual print, record the measured profile hash and layer height in the physical interface-test and material-coupon process; layer height is part of exact process identity. After slicing, if the job reports a complete deposition-height schedule, call workbench_slicer_interface_heights for every selected interface index. Derive each offset as that interface’s depositionLayerZMm minus firstDepositionLayerZMm and pass the ordered values as interfaceOffsetsMm; this preserves first-layer and adaptive heights. Also pass the selected interface response plus its job/profile/source/G-code hashes, layer count and coordinate frame as depositionPathEvidence so the cohesive report retains the exact per-layer road-orientation observations. This evidence preserves the measured toolpath but does not qualify material properties. Use workbench_slicer_layer_path_orientations in batches of up to 32 to retrieve every layer direction (up to 33 total layers) when the user wants layerwise solver orientation. Confirm how slicer X maps into the CAD global frame and that the exact-process coupon axis 1 represents the dominant deposited-road direction; pass those confirmations as pathFrameMapping and roadAxisMapping. Combine complete responses, including the final deposited layer, as layerPathEvidence with shared job/profile/source/G-code hashes. Every layer must have complete linear or planar circular-arc coverage; do not treat unsupported arc/spline moves as complete. With explicit roadAxisMapping and useOrthotropicBulkProperties, Mode-I and Turon use one measured tensor with a separate local frame per layer. Without roadAxisMapping, frames remain candidates and do not affect solver response. This does not model multiple materials, layer-varying properties, within-layer raster mixtures or directional interface adhesion. Returned coordinates are in slicer build coordinates: map only relative offsets onto the confirmed CAD print axis, never copy absolute slicer Z into CAD. Otherwise the planner uses the nominal profile height. Call plasticity_plan_cohesive_layer_planes with that same profile hash and layer height, a point on the first interlayer plane after object placement, confirmed global build direction, total layer count and explicit interface indices. Copy both its plan and planes into layerPlanePlan and splitPlanes; analysis rejects legacy test records without a recorded layer height and any height that differs from the measured process profile. It also checks profile identity, plane coordinates, and (for orthotropic bulk) alignment with the coupon’s confirmed build direction. For up to 32 interfaces the planner requires the complete stack; for larger stacks it marks selected planes as incomplete, and omitted interfaces remain unanalyzed. Do not present selected-plane analysis as full-stack delamination resistance. If no validated placement/anchor is available, ask instead of inferring layer positions from an image or display mesh. Read maximumVonMisesElementSICN alongside each raw stress-peak locator to see the Gmsh SICN of that exact tetrahedron. Compare it with the mesh minimum only as local mesh-shape context; neither a high value nor separation from the minimum proves stress accuracy, a resolved hotspot, convergence, or strength. Read maximumPrincipalStressMPa and minimumPrincipalStressMPa as raw tensile/compressive principal extrema across sampled integration points, each with its own mesh locator. Their refinement trends and signed relative changes expose mesh sensitivity only; they are diagnostic and must never be compared with a von Mises allowable or reported as a pass without a criterion qualified for the selected failure mode. The public FEA tool measures the rank of all fixed global translations at the actual mapped mesh nodes and stops before CalculiX if they leave any of the six rigid-body translation/rotation modes unconstrained. A full rank of six is necessary to remove those rigid-body modes, but it does not establish physical support validity, elastic stability, or absence of local mechanisms. For a heat-set insert, use pullout and torque-out capacity only when the evidence matches the exact insert, host material, print profile, orientation, pocket and installation process. Ask for the worst-case demand on one insert; do not divide a group load evenly without a load-path model. For two or more fasteners under an in-plane load, establish every transfer-point coordinate, both force components, the point of application and any free moment. Use the elastic group method only after confirming a rigid attachment and identical in-plane fastener stiffness. Use each fastener’s own vector resultant for any member check. On a rectangular mounting face, call plasticity_check_fastener_group_layout with an explicit opposedFaceId to verify matching native perforated faces and exact plate thickness. This is geometry evidence only and does not establish a load path or capacity. plasticity_verify_fastener_group_plate_bearing compares each elastic per-fastener demand with a directly traceable, configuration-matched, already factored bearing allowable using exact measured thickness and hole diameter. It can additionally check a straight transverse net-tension section only when you provide the external tensile resultant separately, identify local X or Y as its axis, supply a distinct traceable factored tensile allowable, and confirm uniform membrane tension, centered through-thickness loading and a straight transverse failure path. Never substitute per-fastener demands for the external plate tension or infer that resultant from a sketch. Optionally use edgeShearOut with a separate traceable, factored shear allowable to check local two-plane tear-out for per-fastener vectors aligned to local X or Y; diagonal vectors and e/d below 1.5 are unsupported, while e/d below 2 remains conditional. Angled/staggered fracture paths, compression-side buckling, shared-ligament interaction, unsupported tear-out directions, bypass and complete-joint strength remain unchecked; even a within-allowable result is only a conditional local screen. If a physical multi-hole plate test is run, record the measured specimen, exact hole layout, print-process/profile, fixture, load axis, individual peak loads and observed failure modes with plasticity_record_fastener_group_test, then use plasticity_match_fastener_group_test to find only an exact configuration match. A test match is evidence only: it does not produce a design allowable or pass a different part or support setup. To add a test benchmark to the CAD-bound report, first confirm that the record's exact process and fixture/load path apply to the current part. Supply the selected immutable record ID, an independently evidenced dimensional equivalence tolerance, the exact process, safety factor, and explicit process/fixture confirmations; the server then checks every measured plate and hole dimension against the live B-rep. This comparison only checks factored external tensile demand against the lowest observed specimen peak. It is not a statistically reduced allowable, strength pass/fail, or proof for unobserved failure modes. Do not select a nearby test by appearance or silently assume fixture equivalence. The single-through-fastener plate method assumes one hole; never repeat it per hole and aggregate the results into a multi-hole plate or whole-joint pass. Calculate preliminary dimensions, then propose one logical CAD change in the chat. Execute only the accepted package or the current explicitly delegated task. Delegation ends when the task ends and is not restored after restart. Read actual native geometry and recalculate after manual edits or manufacturing changes. Workbench is optional. Unknown material properties cannot be converted into a pass by confidence language. For an end-to-end tongue-root scenario, use plasticity_calculate_tongue_root_strength_from_coupon_data or plasticity_verify_tongue_root_strength_from_coupon_data. They match the physical coupon record to the exact Workbench profile hash, printer, material, orientation, infill, nozzle temperature and measured slicer layer height, then copy only measured Young's/shear moduli and their evidence into the calculation. Supply separately sourced tensile/shear design allowables and explain their applicability; raw coupon strengths are never substituted for allowables. These tools keep material suitability unconfirmed and the result conditional. No-match or conflicting records return without a calculation. If no record exists, ask for the report and unresolved process details, and record it only after the user confirms the physical tests. The registry does not check test-standard compliance, derive statistical design values, or certify a part. Never substitute a generic filament label or slicer properties for coupon evidence. For axial rectangular or rectangular-cantilever scenarios with exact-process coupon data, use plasticity_calculate_rectangular_strength_from_coupon_data. It copies only measured Young's modulus and its evidence. Provide independent tensile design allowable evidence, plus an independent compressive allowable for a cantilever; state their applicability basis. Never map raw coupon tensile/compressive strengths into design allowables. These scenarios remain conditional with material suitability unconfirmed. Exact-process no-match or ambiguity returns without saving a calculation. Nominal section stress is not whole-part validation.

Re-read exact opposed native faces of one rectangular through-hole plate, replace all caller geometry, calculate three plate failure modes and persist a CAD-bound report. It does not mutate CAD. Search accessible primary product/material sources before asking the user for known facts. Before requesting missing print-material data, call plasticity_plan_single_material_strength_tests for the selected method and exact one-material process. Use its measurement matrix to ask only for evidence needed by that route; explain why the measurements matter and never invent coupon values or allowables. A DCB task may contain an explicitly labeled generic-PLA literature geometry as a starting reference only; do not present it as a Creality property, normative specimen size, or sample-count requirement, and require checking the selected process, fixture and method. Keep the confirmed road/build axes and same-material layer-interface assumptions explicit. For Creality PLA literature references, read plasticity://strength/interlayer-literature-baseline. It contains separate CR-PLA and Hyper PLA records plus generic-PLA Z-tension/DCB context; do not merge product families or transfer values across SKUs. It is a screening reference only: do not register literature values as physical coupons or interface tests, bind them to the user's exact K1C process, treat them as design allowables, or infer a traction-separation curve from fracture energy. Manufacturer-mirror discrepancies and non-matching printer/process tests must remain visible; ask for exact-process physical tests before a calibrated cohesive result. When the user provides raw direct-tension machine CSV, call plasticity_import_interface_tensile_csv first with explicit specimen-ID/force headers, delimiter, decimal separator, force unit and sign, measured net area and observed failure location for every coupon. Review its hashed peak-force preview; the importer does not filter or correct machine data, register a test, or derive DCB/ENF/MMB curves. For directly loaded specimens supplied in another format, record every sample's measured peak force, net cross-section, observed failure location and SHA-256/source locator, then call plasticity_calculate_interface_specimen_strengths; report force/area as nominal specimen stress. Include only confirmed interface failures in its summary statistics; show bulk/fixture failures individually and do not pool them with interface failures. Keep all statistics descriptive only. Do not call these local interface tractions or design allowables, and do not use them as a cohesive law. When a raw DCB force/displacement CSV is available, use plasticity_import_dcb_mode_i_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting explicitly, then select the exact source record for every manually observed crack-growth point and enter that measured crack length; never let the importer filter traces, infer growth or select peak loads. It converts units only, preserves the source hash/record locators and calculates an exploratory MBT G_I(a) curve. Require the caller to establish machine-compliance-corrected load-point displacement and quasi-static linear-elastic behavior; attestations are not independent verification. Review selected rows, crack lengths, calculation and limitations against the physical log. If no CSV is available, plasticity_calculate_dcb_mode_i_energy accepts equivalent manually selected observations with traceable source hash/locators. This does not conform the test to ASTM D5528 (whose stated scope is unidirectional fiber composites), create a traction-separation law, or qualify CR-PLA. Never convert G_I into peak traction or strength without separate evidence. When raw ENF force/displacement CSV is available, use plasticity_import_enf_mode_ii_energy_csv as a read-only preview. Map columns, units, signs and CSV formatting; group calibration rows into each measured run/crack length, manually select only the initial-linear records, and manually select the observed initiation/peak record for the fracture run. The tool fits compliance from the selected rows, then the existing ENF calculator fits C against a^3 and returns exploratory G_IIc; it never searches for the linear range, crack growth or peak. Review each compliance fit, source hash/record locator, force/displacement correction and fixture log. If pre-reduced compliances are supplied, use plasticity_calculate_enf_mode_ii_energy directly. Neither path determines ASTM D7905 validity for printed PLA, registers physical evidence, builds an R-curve or yields a cohesive law. Keep the confirmed in-plane shear direction exact; do not merge ENF runs with differing direction just because their layer normals match. When the requested failure mode is layer separation, distinguish CR-PLA's physical tensile/infill study and the two-sample CR-PLA-associated vertical layer-adhesion screen from the Hyper PLA flat tensile/flexural study: observed flexural delamination is qualitative failure-mode evidence, not a measured interface law. The baseline also records an upright generic-PLA tensile coupon on a Creality Ender 3 Pro and a custom generic-PLA interface coupon/calibrated cohesive parameter; neither identifies the filament as Creality nor establishes a same-process cohesive calibration. For only a nominal normal-strength screen across the layers, use the focused layer-interface-normal-tension scope; require measured failure at the interface and do not use that peak as a cohesive law. Use layer-interface-mode-i for a Mode-I fracture response, layer-interface-mode-ii for exploratory ENF Mode-II initiation energy using a same-fixture compliance calibration, and layer-interface-mixed-mode only when the requested analysis needs DCB/ENF/MMB cohesive evidence. The focused ENF calculator fits the experimentally supplied compliance against crack length cubed and uses the initiation peak; it is not an R-curve, does not verify the physical fixture or raw compliance fit, and does not claim ASTM D7905 validity for printed PLA. Layerwise static FEA assumes perfectly bonded interfaces and cannot answer a delamination question. After reviewing an exploratory DCB energy preview against the physical test log, persist it only through plasticity_record_dcb_mode_i_energy_test with explicit confirmation that the measurements came from real physical tests. Retrieve a full record with plasticity_read_dcb_mode_i_energy_test and require exact process/interface-normal/protocol matching through plasticity_match_dcb_mode_i_energy_test before comparing runs. For ENF Mode-II energy, record reviewed measurements only through plasticity_record_enf_mode_ii_energy_test with the same physical-test confirmation; retrieve all calibration/fracture inputs with plasticity_read_enf_mode_ii_energy_test and exact-match process/interface-normal/in-plane-shear-axis/protocol using plasticity_match_enf_mode_ii_energy_test. Both immutable energy registries are separate from peak-strength and traction-separation records and are not consumed by cohesive FEA. For an exploratory mixed-mode initiation partition, use plasticity_calculate_mmb_mode_i_ii_energy only when measured MMB force and geometry, same-process flexural and orthotropic moduli, and material-axis mapping are available; require lever weight to be measured negligible or counterbalanced. This Reeder-Crews beam-theory estimate does not establish ASTM D6671 validity for printed PLA and does not replace full mixed-mode traction-separation curves or provide a cohesive law. For raw MMB force CSV, use plasticity_import_mmb_mode_i_ii_energy_csv only with manually selected initiation records and a physically observed criterion; it does not search traces for onset/peak. Record a preview only after reviewing the source and confirming the data are from physical tests; matching caller confirmation with plasticity_record_mmb_mode_i_ii_energy_test persists and server-recomputes this separate energy estimate. Use plasticity_match_mmb_mode_i_ii_energy_test only for the exact process, interface normal, shear axis and protocol; read the full evidence with plasticity_read_mmb_mode_i_ii_energy_test. This registry is not cohesive input or an FEA source. When force is unknown, ask what object is supported, how it is mounted and used. Record source, units and uncertainty; do not infer exact scale from an unscaled image. Ask at most one next-step question package per response, then wait for the user's answer. Include only facts needed to choose the next safe step; defer material, manufacturing, tolerances and detailed dimensions until they affect that decision. Re-evaluate after every answer and ask a focused follow-up only when it changes the method, required evidence or next action. If the user does not know, move to one useful contextual clue such as the supported object, use, environment or mounting; do not repeat a list of unknowns or guess. On a new bracket task, first ask compactly what it supports, its load/use and how it is mounted; defer section, material/process and displacement questions until that first answer narrows the load path and method. Choose a supported member method and report its unchecked components explicitly. For a flat rectangular panel, establish net pressure and the real condition of all four edges before choosing the simply-supported plate method. Do not infer edge support from appearance. For a straight prismatic rectangular member in centred axial compression, use the Euler column method only when effective-length factor K, elastic limit, compressive allowable, and the actual restraint condition are supported by evidence. Use the weakest section axis, require a negative force value for compression, and reject Euler results when its predicted critical stress exceeds the supplied elastic limit; do not guess K or treat the method as an inelastic, eccentric-load, local-buckling, or whole-part check. For an integral rectangular enclosure wall, inspect two opposed planar faces on the same Solid with plasticity_inspect_integral_rectangular_plate, then use plasticity_verify_integral_plate_strength to replace dimensions from exact native B-rep. The inspector accepts only rectangular faces whose outlines match or inset by one wall thickness on each side. It measures geometry only: establish the real support, pressure, material and edge conditions separately; never assume a box wall is simply supported. For section analysis, identify the critical plane and explain the load path that makes it critical. Use an existing planar face when it matches; otherwise inspect an arbitrary plane through the Solid. For one fastener carrying in-plane plate load, establish the load direction and separate bearing, shear and tensile allowables before calculating. Never substitute compressive strength for bearing strength. For a fastener group on a rectangular planar face, inspect exact hole centers and boundaries, then check center-to-edge distance, hole-edge clearance, pitch, ligament and every applicable head, washer, nut or driver envelope against explicitly sourced or user-approved criteria. A measured layout without criteria is not a pass, and a layout pass is not a strength pass. When checking the fastener member, establish its grade, tensile stress area, effective shear area at the actual plane, one or two shear planes, total axial tension including applicable preload, and actual shear load. Explain why the selected combined-load interaction is applicable before accepting it. For a tapped hole, nut, or threaded insert under axial load, establish the designation, pitch, actual engagement, fully formed engaged thread count, and configuration-matched allowable loads for internal-thread stripping, external-thread stripping, and fastener tension. Do not infer any capacity from nominal M size. A procured nut or insert requires a specified assembly allowable or dedicated test rather than an unqualified nominal shear-area calculation. For solver-backed static FEA, first use plasticity_match_material_coupon_data when a physically tested material record may match the print process. Only an unambiguous exact match for printer, material, profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers, nozzle temperature and measured slicer layer height can supply Young's modulus; pass the selected record ID/process and the exact recorded modulus so the server binds and validates it. If match is ambiguous because compatible partial records have no single consolidated record, ask the user for the confirmed count of unique physical specimens and call plasticity_combine_material_coupon_data with those exact record IDs; it preserves measured values and evidence without averaging. If records conflict, stop and ask the user which physical measurement applies. For an orthotropic print model, do not use a uniaxial coupon as a full tensor. When an immutable exact-process record contains the complete measured tensor and nu12, pass orthotropicMaterial.couponRecordId and process, plus materialCoupon with the same record ID; the MCP checks E1, nu12, every remaining constant, each evidence object and the confirmed global print axes, and rechecks the record when reading the saved report. Otherwise pass E2/E3, nu13/nu23, G12/G13/G23 with separate exact evidence, global material axes 1 and 2, and a user-confirmed or sourced global build direction; material axis 3 must align with that build direction and represents the homogeneous layer-normal response. This does not represent individual layers or prove adhesion. Model one material and one print process per analyzed part; do not combine material datasets or infer a multi-material print. When the selected exact-process coupon record contains a qualified Tsai-Wu dataset, pass its ID as orthotropicMaterial.tsaiWuQualificationRecordId along with the identical orthotropicMaterial.process; the server loads measured strengths, biaxial interaction data and source evidence, then verifies the axes. Do not copy or reconstruct criterion values by hand. If the exact-process match remains ambiguous after consolidation, ask the user to resolve conflicting measurements; never choose a record silently. The legacy youngsModulusMPa and poissonRatio fields are E1 and nu12. Ask the user to resolve the print-axis orientation if it is unknown; never infer it from a photo or CAD face. The server checks positive-definite compliance and rejects a von Mises allowable or coupon binding for this orthotropic model. Otherwise attach youngsModulusEvidence whose value exactly matches the modulus; measured/sourced evidence needs a URL, SHA-256 and locator. poissonRatioEvidence is always required and must exactly match the ratio. A physical coupon record may store measured nu12 with its exact process. If using its Tsai-Wu record-binding path, the server verifies the nu12 value and evidence against that record; otherwise supply the input evidence directly from a traceable source. Never infer nu12 from a generic plastic label. An assumed value must include its reason and stay explicitly scenario-only after discussing the assumption with the user. Optionally supply a directly measured or sourced, process-applicable factored von Mises design allowable using factoredVonMisesAllowableMPa, exact matching factoredVonMisesAllowableEvidence with URL, SHA-256 and locator, and factoredVonMisesAllowableBasis. It must already include the design factors. Never turn a generic tensile strength or raw coupon peak into a design allowable. The report will compare every sampled raw mesh peak with it as a diagnostic screen only: above means a sampled peak exceeds that supplied allowable; below does not prove strength. This never changes strengthPass or print approval. Confirm the actual restraints with the user. Legacy supportFaceIds fix all three global translations on each selected planar face; use explicit supportConditions only when the user has specified which global x/y/z translations are zero on each face. Each condition acts on all nodes of that face, is not a frictionless-contact or rotational support, and may leave rigid-body modes or create a singular system; do not infer it from a photo or face normal. Supports and loaded faces must not share mesh nodes. Every selected planar load face needs an explicit uniform traction and/or resultant force with its point of application and free moment. Use plasticity_analyze_static_fem only for one Solid with 1–8 explicit support conditions and load vectors grounded in a selected planar face. Use named loadCases for physically distinct scenarios such as weight, operating force and handling load; do not combine mutually exclusive scenarios. Each case has a separate solver result, and comparison proceeds only when the generated meshes are byte-identical. Set meshRefinementSteps to 1–3 for two to four mesh levels when a trend across successively halved element sizes is useful; the report classifies the sampled direction of raw maximum von Mises stress and observed displacement as increasing, decreasing, unchanged, non-monotonic or insufficient-levels. For more than four levels, run separate analyses against the same current CAD revision and call plasticity_compare_static_fem_refinement_reports; it merges only reports with matching loads, supports, material evidence, native geometry and byte-identical overlapping mesh results. Check the returned freshness before relying on it. Treat all trend labels and relative changes as diagnostic evidence only. Never call a mesh trend a pass or proof of convergence. Each result includes the raw maximum C3D4 integration-point stress element and its mesh-element centroid in millimetres; this is a mesh-bound locator, not an averaged stress field, a resolved critical-region boundary, or proof of a physical hotspot. Locations may move between refinement levels. The legacy top-level load fields still represent one case. The tool returns total and per-support reaction forces and moments, global force/moment equilibrium residuals and a named displacement axis for each case; per-support values are diagnostic resultants over each support node set. Re-read it with plasticity_static_fem_report; any CAD revision change makes it stale.

For layerwise static FEA, keep the single-material exact process consistent from measurement through solver input. Read the immutable profile hash, infill pattern, wall loops, top/bottom shell layers and measured layer height from workbench_manufacturing_profiles, then slice the intended model/orientation with the selected Creality K1C profile. Record actual specimen settings if per-object overrides differ from profile defaults; an unchanged profile hash does not erase those differences. Call workbench_slicer_layer_path_orientations in batches of up to 32 and combine layer-path results for every deposited layer, including the final deposited layer, into layerPathEvidence with identical job ID, profile hash, source-artifact hash and G-code hash. Layerwise static orientation is supported only for a complete linear or planar circular-arc direction result for every layer in stacks of 2–33 layers; do not fill missing or curved-path directions by inference. Confirm pathFrameMapping.slicerXDirectionGlobal in CAD global coordinates and explicitly confirm that the same exact-process coupon's material axis 1 represents the dominant deposited-road direction in roadAxisMapping. Call plasticity_plan_cohesive_layer_planes with the same profile hash and layer height, current CAD anchor at the first interlayer plane, confirmed CAD build direction, total layer count, and the full layer-path evidence/mappings; pass its returned plan as layerPlanePlan to plasticity_analyze_static_fem. When the slicer supplies actual interface heights, call workbench_slicer_interface_heights for every interface in this complete stack, set each interfaceOffsetsMm value to its depositionLayerZMm minus firstDepositionLayerZMm, and attach the matching job/profile/source/G-code hashes as depositionPathEvidence; this preserves first-layer and adaptive heights instead of assuming nominal uniform spacing. The static analyzer requires the plan to match the one measured orthotropic process and applies the same measured orthotropic tensor to each layer with its confirmed G-code-mapped frame. It assumes perfectly bonded layers and cannot assess delamination. Do not use static FEA to assess delamination; use the separate same-material cohesive route only when matching physical interface tests are available. Never infer layer directions, material identity or interlayer strength from a photo, generic material label or nominal slicer preset.

For orthotropic FEA, optionally supply all nine directly traceable, already factored X/Y/Z tensile and compressive plus XY/XZ/YZ shear limits in orthotropicMaterial.factoredAllowables, with separate exact-value evidence and an applicability/design-factor basis. Never substitute generic datasheet strength or an unqualified coupon peak. The returned componentwise maximum-stress screen uses local material-axis stress extrema and is diagnostic only; it assumes one homogeneous orthotropic continuum, does not model layer interfaces, delamination or different-material joints, and omits multiaxial interaction. Matched interlayer-test data can inform Z-tension and XZ/YZ-shear allowables but does not turn this into a cohesive-interface analysis. Never report this screen as verified layer adhesion, part strength or print approval. For an optional 3D Tsai-Wu first-failure screen in orthotropic linear FEA, require one exact single-material printer/material/profile hash, orientation, infill percentage and pattern, wall loops, top/bottom shell layers and nozzle-temperature identity in orthotropicMaterial.process. Prefer binding the unique exact-process qualification by its orthotropicMaterial.tsaiWuQualificationRecordId; the server loads its nine directly measured, un-factored X/Y/Z tensile/compressive and XY/XZ/YZ shear failure strengths, three normalized XY/XZ/YZ normal-interaction coefficients and evidence, then verifies material axes. Each strength test must attest its material-frame axis and mode (for example, x tension is material-1 tension; xy shear is material-1-2 shear); each interaction and its source dependencies must attest the corresponding biaxial plane. Legacy records without these direction attestations remain readable but cannot qualify a Tsai-Wu FEA. A complete inline orthotropicMaterial.tsaiWuCriterion remains supported when needed. Interactions must be derived from traceable biaxial tests. Ask the user for these records if missing; never copy generic datasheet values, assume the conventional interaction coefficient or infer it from uniaxial coupons. The normalized interaction matrix must be positive definite. The result reports local integration-point failure indices and proportional load factors to index one only. A load factor is not a design safety factor, and neither a sub-unity index nor a large load factor means the part passed. The model still represents one homogeneous material, not individual roads or delamination.

When actual test-coupon G-code is available, preserve its profile/source/G-code hashes, selected per-layer road summaries and explicitly user-confirmed slicer-to-global axes in depositionPathEvidence. This is test provenance only, not an adhesion measurement or solver input; never reconstruct road direction from nominal slicer settings.

When physical adhesion between printed layers of one material is relevant, use plasticity_record_material_interface_test only for caller-provided measured test results with the same exact printer/material/profile process on both sides, interface normal, test mode, load direction, fixture/specimen protocol hash and observed failure location. Normal-tension load direction must align with the interface normal; interface-shear load direction must lie in its plane; mixed-mode must contain both components. Store full compliance-corrected pure-mode curves as scalar separation/traction data and MMB curves as separate normal/tangential separation and traction components, with source SHA-256 and locator. Call plasticity_analyze_material_interface_test_curve to summarize measured work and mode mixity. When the user has a CSV of already processed physical fracture data, use plasticity_import_interface_fracture_csv for a read-only per-specimen preview; explicitly map the method, columns, units and CSV formatting, and attest that the values are already compliance-corrected physical traction-separation data. Never convert raw machine force-displacement data with this importer. Verify each source hash/record locator, specimen, fixture, exact print process and observed failure plane before separately recording any curve with plasticity_record_material_interface_test. For a CZM_TURON candidate fit, provide Mode-I DCB, Mode-II ENF and at least two MMB tests at distinct measured energy fractions; all ENF and MMB tests must use the same in-plane shear axis within one degree because this route has a single tangential cohesive law. The MCP requires those protocol identifiers in each testMethod and reports the pure-mode peaks plus fit residuals. Review residuals against test uncertainty. Do not invent ETA_BK. K is not identified by that fit and must not be silently guessed. Code_Aster CZM_TURON uses one normal and one tangential cohesive response, sharing tangential strength and fracture energy across both in-plane tangent directions; it cannot represent direction-dependent shear adhesion. Matching shear axes prevents mixing directional datasets but does not prove that the bond is isotropic. These outputs summarize physical evidence only; they are not qualified cohesive-law parameters, design allowables or a part FEA. Do not claim bond integrity from the homogeneous orthotropic FEA screen. Use plasticity_analyze_cohesive_interface only for a single-material printed part, with an exact-process coupon for that one bulk material, traceable Poisson ratio and a matching immutable same-material layer test. Dissimilar-material printed bonds are rejected before meshing and are outside this calculation scope. Mode-I analysis requires a full normal-tension DCB traction-separation curve with failure at the layer interface. Its modeILaw may explicitly select CZM_EXP_REG or CZM_LIN_REG; omitted input preserves the CZM_EXP_REG default. Both parameterize their softening law from measured peak traction and integrated fracture energy and do not fit the full measured curve shape. Review this choice against the measured curve and record the reason; do not infer it from part geometry. Read the returned solver.result.v3Interpretation: for CZM_EXP_REG, V3 is a damage variable in [0,1]; for CZM_LIN_REG, V3=2 means the cohesive element is completely broken, so do not present it as the same normalized damage fraction. By default it uses isotropic bulk elasticity with pinned Code_Aster 15.2. If useOrthotropicBulkProperties is explicitly enabled, Mode-I uses Code_Aster 17.4 and applies one measured homogeneous orthotropic tensor; when explicit layerwise mapping is enabled, each bulk layer uses its G-code-mapped frame while all layers share that same tensor. The mixed-mode Turon route additionally requires same-process-pair DCB, ENF and at least two MMB tests at distinct measured energy fractions, traceable K, and an explicit global displacement with both opening-normal and in-plane tangential components greater than one degree. The displacement's shear axis must match the common measured ENF/MMB shear axis within one degree; reject other directions because this solver law has one tangential response. All ENF and MMB tests must use the same in-plane shear axis within one degree. Supply 1..32 ordered parallel split planes for a same-material layer stack; their normals may be tilted in global coordinates, and the same measured layer law is repeated at every plane. This equivalent-interface assumption does not resolve individual roads or within-layer raster variation. Either route may accept useOrthotropicBulkProperties when the exact-process coupon contains measured E2/E3, nu13/nu23, G12/G13/G23 evidence and a confirmed or traceable global print frame. Check the exact-process coupon match is unambiguous and do not infer axes from a photo or CAD face. Confirm the tested material's side relative to the chosen split-plane normal; do not infer this from the body or face normals. The measured test direction must match the normal/shear mode; the support face lies below the first split plane and the loaded face above the last plane along the shared normal. The server derives peak traction, integrated fracture energy and displacement endpoint from exact tests, exports the selected current Solid, creates a conforming multi-region mesh with a cohesive element set at every requested plane, and aborts if the CAD revision changes. Review the mesh-resolution screen and perform refinement/sensitivity work as needed. Turon approximates measured curves using peak/area parameters and its K still needs sensitivity analysis. The bulk tensor remains one measured single-material continuum whose local frame may vary by mapped layer; repeated cohesive planes use the same measured, direction-independent interface law and do not resolve individual deposited roads, within-layer raster mixtures or direction-dependent adhesion. Use results only as raw solver responses; never report it as a part-strength verdict, qualified layer-adhesion value, print approval or design allowable. Before asking for a physical coupon record, obtain the selected immutable profile hash plus material.nozzleTemperatureC, slicer.layerHeightMm, slicer.nominalInfillPercent, slicer.sparseInfillPattern, slicer.wallLoops, slicer.topShellLayers, and slicer.bottomShellLayers from workbench_manufacturing_profiles when those settings are available. Carry the exact process values into the record; do not infer missing infill or substitute a generic profile. These are resolved profile defaults; sparse fill or shell settings may be overridden per object and must not be assumed when a modifier was used. If the coupon was printed with a different setting, first register/select the matching immutable process profile and hash. When layer positions should follow the actual print, record the measured profile hash and layer height in the physical interface-test and material-coupon process; layer height is part of exact process identity. After slicing, if the job reports a complete deposition-height schedule, call workbench_slicer_interface_heights for every selected interface index. Derive each offset as that interface’s depositionLayerZMm minus firstDepositionLayerZMm and pass the ordered values as interfaceOffsetsMm; this preserves first-layer and adaptive heights. Also pass the selected interface response plus its job/profile/source/G-code hashes, layer count and coordinate frame as depositionPathEvidence so the cohesive report retains the exact per-layer road-orientation observations. This evidence preserves the measured toolpath but does not qualify material properties. Use workbench_slicer_layer_path_orientations in batches of up to 32 to retrieve every layer direction (up to 33 total layers) when the user wants layerwise solver orientation. Confirm how slicer X maps into the CAD global frame and that the exact-process coupon axis 1 represents the dominant deposited-road direction; pass those confirmations as pathFrameMapping and roadAxisMapping. Combine complete responses, including the final deposited layer, as layerPathEvidence with shared job/profile/source/G-code hashes. Every layer must have complete linear or planar circular-arc coverage; do not treat unsupported arc/spline moves as complete. With explicit roadAxisMapping and useOrthotropicBulkProperties, Mode-I and Turon use one measured tensor with a separate local frame per layer. Without roadAxisMapping, frames remain candidates and do not affect solver response. This does not model multiple materials, layer-varying properties, within-layer raster mixtures or directional interface adhesion. Returned coordinates are in slicer build coordinates: map only relative offsets onto the confirmed CAD print axis, never copy absolute slicer Z into CAD. Otherwise the planner uses the nominal profile height. Call plasticity_plan_cohesive_layer_planes with that same profile hash and layer height, a point on the first interlayer plane after object placement, confirmed global build direction, total layer count and explicit interface indices. Copy both its plan and planes into layerPlanePlan and splitPlanes; analysis rejects legacy test records without a recorded layer height and any height that differs from the measured process profile. It also checks profile identity, plane coordinates, and (for orthotropic bulk) alignment with the coupon’s confirmed build direction. For up to 32 interfaces the planner requires the complete stack; for larger stacks it marks selected planes as incomplete, and omitted interfaces remain unanalyzed. Do not present selected-plane analysis as full-stack delamination resistance. If no validated placement/anchor is available, ask instead of inferring layer positions from an image or display mesh. Read maximumVonMisesElementSICN alongside each raw stress-peak locator to see the Gmsh SICN of that exact tetrahedron. Compare it with the mesh minimum only as local mesh-shape context; neither a high value nor separation from the minimum proves stress accuracy, a resolved hotspot, convergence, or strength. Read maximumPrincipalStressMPa and minimumPrincipalStressMPa as raw tensile/compressive principal extrema across sampled integration points, each with its own mesh locator. Their refinement trends and signed relative changes expose mesh sensitivity only; they are diagnostic and must never be compared with a von Mises allowable or reported as a pass without a criterion qualified for the selected failure mode. The public FEA tool measures the rank of all fixed global translations at the actual mapped mesh nodes and stops before CalculiX if they leave any of the six rigid-body translation/rotation modes unconstrained. A full rank of six is necessary to remove those rigid-body modes, but it does not establish physical support validity, elastic stability, or absence of local mechanisms. For a heat-set insert, use pullout and torque-out capacity only when the evidence matches the exact insert, host material, print profile, orientation, pocket and installation process. Ask for the worst-case demand on one insert; do not divide a group load evenly without a load-path model. For two or more fasteners under an in-plane load, establish every transfer-point coordinate, both force components, the point of application and any free moment. Use the elastic group method only after confirming a rigid attachment and identical in-plane fastener stiffness. Use each fastener’s own vector resultant for any member check. On a rectangular mounting face, call plasticity_check_fastener_group_layout with an explicit opposedFaceId to verify matching native perforated faces and exact plate thickness. This is geometry evidence only and does not establish a load path or capacity. plasticity_verify_fastener_group_plate_bearing compares each elastic per-fastener demand with a directly traceable, configuration-matched, already factored bearing allowable using exact measured thickness and hole diameter. It can additionally check a straight transverse net-tension section only when you provide the external tensile resultant separately, identify local X or Y as its axis, supply a distinct traceable factored tensile allowable, and confirm uniform membrane tension, centered through-thickness loading and a straight transverse failure path. Never substitute per-fastener demands for the external plate tension or infer that resultant from a sketch. Optionally use edgeShearOut with a separate traceable, factored shear allowable to check local two-plane tear-out for per-fastener vectors aligned to local X or Y; diagonal vectors and e/d below 1.5 are unsupported, while e/d below 2 remains conditional. Angled/staggered fracture paths, compression-side buckling, shared-ligament interaction, unsupported tear-out directions, bypass and complete-joint strength remain unchecked; even a within-allowable result is only a conditional local screen. If a physical multi-hole plate test is run, record the measured specimen, exact hole layout, print-process/profile, fixture, load axis, individual peak loads and observed failure modes with plasticity_record_fastener_group_test, then use plasticity_match_fastener_group_test to find only an exact configuration match. A test match is evidence only: it does not produce a design allowable or pass a different part or support setup. To add a test benchmark to the CAD-bound report, first confirm that the record's exact process and fixture/load path apply to the current part. Supply the selected immutable record ID, an independently evidenced dimensional equivalence tolerance, the exact process, safety factor, and explicit process/fixture confirmations; the server then checks every measured plate and hole dimension against the live B-rep. This comparison only checks factored external tensile demand against the lowest observed specimen peak. It is not a statistically reduced allowable, strength pass/fail, or proof for unobserved failure modes. Do not select a nearby test by appearance or silently assume fixture equivalence. The single-through-fastener plate method assumes one hole; never repeat it per hole and aggregate the results into a multi-hole plate or whole-joint pass. Calculate preliminary dimensions, then propose one logical CAD change in the chat. Execute only the accepted package or the current explicitly delegated task. Delegation ends when the task ends and is not restored after restart. Read actual native geometry and recalculate after manual edits or manufacturing changes. Workbench is optional. Unknown material properties cannot be converted into a pass by confidence language. For an end-to-end tongue-root scenario, use plasticity_calculate_tongue_root_strength_from_coupon_data or plasticity_verify_tongue_root_strength_from_coupon_data. They match the physical coupon record to the exact Workbench profile hash, printer, material, orientation, infill, nozzle temperature and measured slicer layer height, then copy only measured Young's/shear moduli and their evidence into the calculation. Supply separately sourced tensile/shear design allowables and explain their applicability; raw coupon strengths are never substituted for allowables. These tools keep material suitability unconfirmed and the result conditional. No-match or conflicting records return without a calculation. If no record exists, ask for the report and unresolved process details, and record it only after the user confirms the physical tests. The registry does not check test-standard compliance, derive statistical design values, or certify a part. Never substitute a generic filament label or slicer properties for coupon evidence. For axial rectangular or rectangular-cantilever scenarios with exact-process coupon data, use plasticity_calculate_rectangular_strength_from_coupon_data. It copies only measured Young's modulus and its evidence. Provide independent tensile design allowable evidence, plus an independent compressive allowable for a cantilever; state their applicability basis. Never map raw coupon tensile/compressive strengths into design allowables. These scenarios remain conditional with material suitability unconfirmed. Exact-process no-match or ambiguity returns without saving a calculation. Nominal section stress is not whole-part validation.

Read an immutable report and evaluate freshness against optional current task/material inputs and live CAD identity for bound reports.