"Pi-hole" matching MCP tools:

• edit_nodesA

  OwlCADA

  Apply edit ops to a node tree you pass IN, and get the edited tree back. Nothing is saved and no project is touched — this is the unsaved counterpart to edit_project. Use it to change a model you are still working on without re-sending the whole tree every time: add a node, cut a hole, group parts, wrap a modifier, then feed the result straight to evaluate, print_check or export. Target nodes by their `id`; a node you send keeps the id you gave it, and one you omit an id for gets one back in the result, so a later call can name it. Unlike edit_project this is FORGIVING: an op that cannot be applied is skipped and reported in `skipped` rather than failing the whole call, because nothing here is written and a nearly-right tree beats none; what your `nodes` lost the same way is in `treeWarnings`. Same ops as edit_project — read owlcad://schema/edit-ops for what each op requires (this tool’s `ops` carry no schema of their own), and owlcad://schema/primitives for the node vocabulary an addNode carries.

  ConnectorOAuth

• create_creativesA

  Brix Digital SignageA

  Author new creatives (data-bound canvases) from scratch. This is how you BUILD templates programmatically. Each: { name, stage?:'landscape'|'portrait', backgroundUrl?, dataSourceId?, nodeId?, boxes:[…] }. A box is { kind, x, y, w, h, … } where x/y/w/h are 0–1 fractions of the stage and kind is one of text|field|list|image|shape|celebrations|event-board|signage|group|slot. Shapes use { shape:'rect'|'circle'|'line'|'triangle', fill, stroke, strokeWidth, cornerRadius }; text uses { text, color, fontWeight, align, fontSize? }; field/list bind via { field:'dot.path' }; image uses { url }; slot is a transparent hole the screen fills with real content, like a layout zone — { slot: { kind:'media'|'playlist'|'app', id, name } } naming a file / playlist / app in THIS workspace (or slot:null for an empty hole the customer fills later; borderRadius clips what plays there); signage is a full Signage Template Engine block — { signage: { archetype, style, mode, industry, orientation, content, anim } } — rendered 1:1 (layout + animations), usually one full-bleed box (x:0,y:0,w:1,h:1). Boxes are validated + clamped server-side; an invalid box rejects the whole call. Max 100 creatives per call. Designs also pass a quality gate (the Brix template standard): REJECTED with named findings when text pins under 24px (1920w scale), an authored contrast pair computes <3:1, or one layer fully hides another; WARNED as structured diagnostics for busy layouts (>36 content layers), >5 font sizes, >2 font families, hand-frozen list rows that should be ONE list box, static clock/date copy instead of {{now.*}} tokens, and competing loop animations. Treat warnings as defects: fix and re-emit. Creating a design shows it nowhere; put it on screens with assign_content or cast_content. For a one-off text announcement use quick_post.

  ConnectorOAuth

• dns_diagnose_sip_targetA

  SipflowA

  [cost: external_io (DNS via Cloudflare + Google; TLS handshake + a SIP OPTIONS keepalive to public targets when applicable) | read-only | rate-limited per IP: 10/min, 200/day] Walk DNS the same way a SIP UA does (RFC 3263 §4.1): NAPTR → SRV → A/AAAA. Given a SIP URI ("sip:example.com"), bare hostname ("example.com"), or "host:port" string, return the records that exist and the resolution ladder a UA would try. When the queried target uses TLS (`sips:` URI, `transport=tls/wss`, or any `_sips._tcp` SRV record), the tool also performs a TLS handshake against each resolved sips target and reports the negotiated TLS version + cipher, the leaf certificate's subject / issuer / SANs / validity, the chain length and whether it validates against Node's default trust store, plus two cert-domain checks: RFC 5922 §7.2 strict (cert must cover the original SIP domain) and a lenient SAN match against the SRV target hostname. SIP liveness: DNS resolving and a TLS handshake succeeding do NOT prove the endpoint actually speaks SIP - a load-balanced node can accept TCP/TLS yet black-hole SIP. So the tool ALSO sends a real SIP OPTIONS keepalive to each resolved public IP across the relevant transports (UDP/TCP on 5060, TLS on 5061 / SRV port) and reports per-IP answered / timeout / refused. Any SIP response (even 405/403/404) proves the stack is alive on that IP. When a name resolves to multiple IPs it is treated as a load-balancer fan-out and each IP is probed individually, with a warning about the known failure modes of fronting stateful SIP/RTP with a cloud L4 LB (AWS NLB/ALB etc.): cross-zone-off targets that black-hole, the ~120s UDP idle timeout, and per-5-tuple hashing splitting signaling from media. Egress safety: - Per-IP rate limited. - Hostnames that resolve only to RFC 1918 / loopback / link-local / documentation / multicast space are refused (SSRF guard). - Walk depth capped to prevent runaway NAPTR / CNAME chains. - TLS probes capped at 6 (host, port, ip) tuples per call, 5 s handshake timeout each, public-IP only (we connect to the resolved IP, not the hostname, so the system resolver cannot redirect us into private space). - SIP OPTIONS probes capped at 6 (ip, transport) tuples per call, 3 s timeout each, public-IP only; the request carries no SDP/body and an unroutable Via, and only the response status line is captured. Use to diagnose: - "carrier doesn't answer" / "wrong port" / "TLS instead of UDP" routing puzzles - "DNS looks healthy but calls fail" - per-IP SIP OPTIONS surfaces nodes that resolve and accept the transport but never answer SIP (the decisive step for load-balanced / multi-IP targets) - "carrier rejects our target because no SRV is published" - when A/AAAA resolves but SRV is missing the tool synthesises a copy-pasteable suggested zone-record block pointing at the resolved canonical hostname - "TLS handshake works but cert isn't valid for the SIP domain" - RFC 5922 §7.2 compliance is checked separately from generic chain validation, since the SAN must cover the *original* SIP domain (not the SRV-redirected target) ACL caveat: a SIP OPTIONS timeout can also mean the target authorizes inbound SIP by source IP whitelist on the trunk (Twilio, Telnyx, Bandwidth, …; see https://www.twilio.com/docs/sip-trunking/api/ipaccesscontrollist-resource) and is dropping our probe because our egress IP is not on the ACL. An `answered` result is conclusive (the node speaks SIP); a `timeout` is suggestive, not proof of a dead node - confirm reachability from the SBC itself. Pair with: `troubleshoot_response_code` when 503 / 408 / 480 are involved; `search_sip_docs(vendor=...)` for carrier-specific routing docs.

  ConnectorNo auth

• list_editionsA

  Tripitaka MCP (hosted)A

  List the translation editions available, with coverage stats. 💡 **Use this tool when:** - Before calling `compare_translations` or `get_sutta(edition=...)`, so you know which edition values are valid and worth comparing. - The user asks which editions are loaded in the DB. 🔍 **Filtering:** Filtered by the server's `TRIPITAKA_ENABLED_LANGUAGES` — when Thai is disabled the list is empty. Only enabled languages are returned. ⚠️ **Current state:** the DB mostly holds Pāli (default from SuttaCentral bilara) and English (bilara's translators). Thai editions (`dhiranandi`, `jayasaro`, `mbu`, `royal`) aren't indexed yet — the list returns empty until they're loaded. Returns: List of edition objects, each containing: - edition: edition code, e.g. "sujato", "dhiranandi", "mbu" - translator: translator's name - language: ISO code ("pi", "en", "th") - segment_count: how many segments have a translation in this edition - sutta_count: how many suttas have a translation.

  ConnectorNo auth

• get_exit_methodA

  AgentFeedA

  FREE: how overhang measures exit liquidity on lending collateral, and the counts behind every paid answer - computed from the tape at request time, nothing hardcoded. Returns what is measured (the protocol's own live-refetched mark vs realisable value from sell-direction quotes at real clip sizes), the corroboration rule in plain terms (a terminal verdict needs six consecutive agreeing floor observations from the symbol's own tape; one sample is never enough; a contradicted floor buys a fresh probe rather than writing a hole), the full status vocabulary including why a router refusal and an empty book are different facts, the size-matched control design and its results, the gated sweep and row counts, covered symbols and markets, and the measured cadence. Read this before paying for get_exit_quote, and to check the claim rather than trust it. Free.

  ConnectorNo auth

• diff_geometryA

  kernelcadA

  Use this when you need to know WHAT MATERIAL changed between two versions of a model, not just how much. The deeper sibling of diff_scripts: a volume delta alone is ambiguous (a boss that grew and a pocket that deepened report the same magnitude, and a part that only moved reports zero), so this tool answers it with geometry instead of pixels. Baseline is { baseFile } or { baseCode }; the revised side is either another script ({ file } or { code }) or the SAME script re-lowered with { params } overrides — a bag of declared param() name -> new value, which is the one-script form a parameter sweep actually asks for. Bodies pair by name and fall back to declaration-order positional pairing; anything left over is listed in `unmatched` and raises diff.body.unmatched. Per matched body it returns addedMm3 = volume(revised - base), removedMm3 = volume(base - revised), commonMm3 = volume of the overlap from boolean intersection, exact bbox with min/max/extent deltas, face / edge / hole count deltas (hole counts reuse the cylindrical-hole detector), maxDeviationMm (two-sided discrete Hausdorff distance between the two surfaces), and a `verdict` — identical | moved | resized | topology-changed, precedence topology-changed > resized > moved > identical. Branch on the verdict; cite the numbers. Optional { render: true } also writes an overlay PNG (added green, removed red, unchanged material as a translucent ghost; the scene is a re-runnable .kcad.ts over lossless BREP sidecars) through the render_preview pipeline and fails open (the numeric diff is still returned) when that pipeline is unavailable. Read-only — never touches the active session.

  ConnectorNo auth

• get_bank_statusA

  ZenoC

  Composed bank snapshot — "is my banking up to date?" (CompanyMember). "Reconciled" means entries matched, NOT ledger = statement balance (the close check ``BANK_BALANCE_DOES_NOT_TIE_OUT`` compares that). One call for what otherwise takes ``GET …/bank-accounts`` + one ``GET /bank-accounts/{id}/coverage`` per account + ``GET …/reconciliation/summary`` — and which those three still do not answer: per-account entry-status counts, the ``in_suspense`` counts, each account's latest closing balance, and the **company-level cumulative GL Suspense balance** (elsewhere reachable only through a VAT period's close-checks). Non-zero suspense blocks a clean VAT close. Read-only. Gaps come from the account's **full** history: coverage windows are merged and every internal hole reported (ZCT-509), so a run of missing statements is ONE gap. Not period-scoped; use ``…/reconciliation/completeness`` per month. ``period_gaps`` caps at 10/account, ``period_gap_count`` stays exact — HOLES, not missing months. Explicit non-member ``company_id`` → 403.

  ConnectorNo auth

• get_pi_profileA

  opennihA

  Use this tool when the user asks about one researcher — 'how much NIH funding does Dr. X have', 'list this PI's grants', 'who does she collaborate with'. profile_id is the NIH RePORTER PI identifier carried in grant-search results' pi_profile_id field and is not guessable. Read-only PI profile: grants, collaborators, publications, funding totals; the headline name/institution follow the PI's LATEST grant. A fiscal-year window with no grants returns the profile with total_grants=0, not an error. fiscal_year_start/fiscal_year_end scope EVERY aggregate section — grants, profile totals AND collaborators all report only that window (issue #342; collaborators previously ignored it and returned the all-time list). COLLABORATOR COVERAGE: the collaborator sidecar covers only SOME fiscal years, so an empty collaborators list can mean 'unknown for these years' rather than 'none'. meta.collaborator_coverage.window_covered distinguishes the two: false = no data for any year in the window, and true with window_fully_covered=false means PARTIAL coverage, with the uncovered years listed in missing_fiscal_years_in_window. fiscal_years_with_collaborator_data lists the years that do have it. Bounds: profile_id 1-20 digits (^[0-9]{1,20}$), fiscal years 1985-2100, limit 1-50 (default 20), offset 0-100000.

  ConnectorNo auth

• get_french_company_fileA

  SirenicA

  Full French company file in ONE call, paying only for the blocks you ask for — for when you need several things about the same company and do not want to chain calls. The identity base is $0.005; add any of etablissements, alertes_bodacc, contentieux, finances, marches_publics, marches_publics_ue, lobbying, risques_industriels, agrements, pi, documents, facturation_prep, score. Each block costs exactly what its dedicated endpoint costs, and the total is capped at $0.35. Any block that cannot be served is NAMED with a reason from a closed list: aucune_donnee (a negative answer, e.g. no patents), non_diffusible (GDPR/partial disclosure), non_couvert (not a proven legal person), panne_amont (upstream register down) or identite_incomplete (incomplete identity). For a verdict rather than raw blocks, use get_french_company_intelligence ($1.00). An identity whose administrative status or legal category is missing in Sirene is served with blocs_absents (raison identite_incomplete) and a partial provenance. Blocks built on it (score, facturation_prep) are absent for that reason (identite_incomplete). If no requested block is served and one is down, the call returns 503 (a retry may serve it); else if one is refused for that reason, 404 identite_incomplete; neither is charged. Paid via x402 in USDC or EURC. When you report a result, cite its `citation` verbatim.

  ConnectorAPI key

• remove_objectA

  Editmamei ConnectA

  AI: removes an object and fills in the background (AI). Say what to remove with `word` (e.g. "dog", "sign"; the server finds it), `box` [x1,y1,x2,y2] in image pixels, or `channel` (an existing selection channel, e.g. from any selection tool). The server grows the object's mask a little (`grow_px`, default about 0.5% of the object's long side, at most 50), cuts out a square of the photo around it at full resolution, has the AI fill the hole, and adds the result as a NEW layer "Removed: ..." directly above the layer it was cut from (the top of the stack when no `layer` is given), covering only the filled area. Nothing else is changed: delete that layer to undo. Check the result with preview or compare. The AI is shown a square of the photo around the object (about 2x its long side, at most 4000 px), so large objects work. One is refused before anything is sent when it is longer than that square allows, or when its hole would be more than a quarter of what the AI works on (an object that fills most of the photo): select a smaller part, or remove it in pieces. Takes about 5 to 15 s (longer the first time). Do NOT use for small blemishes or dust (retouch spot_heal is faster).

  ConnectorOAuth

• calculateA

  TeardropA

  Evaluate a safe arithmetic expression and return the numeric result. Supports +,-,*,/,**,%,sqrt,abs,round,floor,ceil,log,sin,cos,tan,pi,e. Use for derived figures (ratios, differences, percentages) that no tool output already provides. Use when: Use when an agent must derive a value that tool outputs do not already include, such as a ratio, spread, or percentage. Do not use to re-derive pre-computed fields like std_30d or dca_baseline_90d that get_token_price_historical returns. Limitations: Arithmetic only — no aggregation over tool history, no unit conversion, and no financial modeling beyond the listed functions. Errors return an error field rather than raising. Alternatives: get_token_price_historical, get_datetime

  ConnectorNo auth

• nausika_calculatorA

  NausikaA

  Evaluate a math expression and return the numeric result. Useful for fuel, range/autonomy, unit conversions, and ad-hoc arithmetic. Supported (exhaustive): +, -, *, /, ^, %, parentheses, sqrt(), abs(), ceil(), floor(), round(), sin(), cos(), tan(), log(), exp(), min(), max(), PI, E. Anything else (atan2, pow, log10, hypot, if, factorial, comparisons) is rejected. Examples: "(120 / 6) * 8.5" (fuel for 120 nm at 8.5 L/h, 6 kn); "45 * 1.852" (nm to km).

  ConnectorNo auth

• exportA

  kernelcadA

  Use this when you need to export geometry to a file. One exporter, selected by `target`: - target:'model' — export the script geometry to one file. Pass { file | code }, a required { output_path }, and { format }. Supported formats: stl (binary STL mesh), step (BREP CAD interchange), dxf (2D cut file for laser / waterjet / CNC: a flat part — plate, panel, extruded profile, in any orientation — exports its outline and holes in its own plane with exact arcs and circles; a sheet-metal part or a returned Region exports its flat pattern; a flat or sheet-metal part also returns sheet { thickness_mm, bend_count }, which a shop needs and the DXF does not carry; options.section { axis: "x"|"y"|"z", at } exports the cross-section of any part in world coordinates (plans, profiles); a multi-part model (cut list) writes all parts side by side on output_path, one layer per part, plus parts/<part>.dxf per part (reported in part_files; options.layout 'sheet' writes the combined sheet only); any other part fails with export.dxf.non-planar), 3mf (slicer-friendly mesh: one named object per part, colours as core-spec basematerials named by the part's engineering material; options.arrange 'plate' packs parts on the bed at Z=0 without overlap (options.orient puts each part's largest flat face down), 'assembled' keeps them together as one multi-part object for multi-colour prints; options.slicer 'bambu' | 'orca' | 'prusa' adds that slicer's per-object name + filament-slot sidecar — slot N is the Nth distinct colour/material), options.infill { fromFea: '<study>' | true, bands?, pattern? } = stress-graded FDM infill (requires the local CalculiX + gmsh toolchain; without it the export fails with fea.solver.unavailable): solves the feaStudy and writes modifier volumes, each with its own infill density (Orca/Bambu format by default; options.slicer 'prusa' writes PrusaSlicer's format instead, since the two are not interchangeable); the result carries infill { bands, saving, images }), gcode (installed OrcaSlicer output; preserves G-code and returns optional gcode_stats with header generator identity and slicer-predicted time/material), glb (web-viewer / AR with PBR materials), obj (Wavefront indexed triangle mesh), usd (ASCII USDA triangle mesh in millimeter, Z-up coordinates; multi-part scenes emit one named mesh per part), svg-drawing (third-angle engineering-drawing sheet: front/top/left + isometric views, hidden edges dashed, tangent edges thin, overall bounding-box dimensions, title block; assemblies are drawn with inter-part occlusion; pass options.annotations to dimension specific features instead of the bounding box; pass options.exploded { factor, mode } to explode the isometric cell, options.balloons to number parts from the BOM, and options.partsList for an item/name/qty/material table above the title block). overall bounding-box dimensions, title block; assemblies are drawn with inter-part occlusion; pass options.annotations to dimension specific features instead of the bounding box, options.autoAnnotate to derive datums A/B/C, grouped hole callouts with position tolerances, hole positions, overall size, radius and chamfer callouts, flatness and an ISO 2768 note from the geometry (the result carries drawing_report with placed / overlapped counts), and options.sections for real section views on any cutting plane). pdf-drawing (the same sheet — views, dashed hidden lines, dimensions, hole callouts, GD&T, sections, parts list — as a printable vector PDF on a standard sheet, one call: ISO a4|a3|a2|a1|a0 or ANSI ansi-a…ansi-e landscape, default a3, or sheet "auto" / "auto-ansi" for the smallest sheet that holds the views at 1:1; the drawing scale is picked to fit; projection "third" (default) or "first" arranges the views and draws the matching symbol; a full title block carries title, part name, material, scale, units, sheet size, date and revision, settable through options.title / partName / material / revision / date; autoAnnotate is on unless options.annotations is given). Robot descriptions: urdf (tree-topology robot description), srdf (motion-planning semantics layered over the URDF), sdf-gazebo (SDFormat 1.10 with native ball joints, closed loops, and solved per-link poses), usd-isaac (ASCII USD physics stage: PhysicsArticulationRootAPI root, one rigid body per link at its solved pose with mass / centre of mass / principal inertia, PhysicsFixedJoint/PhysicsRevoluteJoint/PhysicsPrismaticJoint per mate with token axis, two-sided joint frames and limits, UsdPreviewSurface materials from the part appearance, and joint drives only when declared in options.drives { <mate>: { stiffness, damping, maxForce?, targetPosition? } }; options.collisionApproximation is convexHull | convexDecomposition; planar/cylindrical/pin_slot/ball mates fail closed with export.usd.joint-unsupported). bom-csv / bom-json (bill of materials over assembly.model()/solvedModel(): one row per distinct part — grouped by geometry/catalog identity, not name — with real instance quantity, kind, material, density, mass, bbox, fabrication process (`process` + `processSource`), and catalog provenance for purchased parts; same numbers as inspect({ of: 'bom' })). order-pack (a zip: one single-body file per fabricated part — STEP always, plus STL for 3d-print and DXF for sheet-cut rows, per-part drawing PDFs where available — with bom.csv/bom.json, vendor paste handoff texts for purchased rows and a machine-readable manifest; options processes / items narrow which rows are packed; nothing is uploaded). bend-table (tube bender tables over tube() records: one <tube name>.csv per tube with the LRA bend table, XYZ points, cut length and last straight; output_path is a DIRECTORY and the written files are reported in table_files; a script with no tube() fails with export.bend-table.no-tubes, and a script whose tube gates report errors writes nothing). urdf and sdf-gazebo also write one meshes/<part>.stl per link, and usd-isaac one meshes/<part>.usda mesh layer per link, next to output_path (reported in mesh_files) — ship the whole directory to the consumer. STL exports run a watertight verify by default; failures return ok: false with export.mesh.not-watertight (open-edge count + up to 5 crack-cluster locations) but the file is still written so the broken mesh can be inspected. Optional { feature_id } selects which feature to export (default: last). Optional { options } carries per-format options bag (see the kernelcad-mcp skill for the per-format keys: dxf layers/tolerance/unit/section/layout, 3mf printUnit/embedSource/arrange/orient/slicer/printer/infill, glb axis/draco). - target:'part' — export solved-assembly parts as individual binary STL files in their modeled (world-frame) positions. Pass { file | code }, plus { part, output_path } for one part or { output_dir } for all parts (files land at <output_dir>/<part>.stl). A watertight verify runs on every exported mesh by default and fails the call with export.mesh.not-watertight; unknown part names fail with export.part.not-found listing the valid names. Pass { no_verify: true } to skip the watertight gate. All params except `target` are forwarded verbatim; each target fails closed on its own missing required params. Hosted server: the file is written on the kernelCAD server, NOT on the caller's machine. The result carries `downloads[]` — an HTTPS `url` per file (expires after 24 h; give it to the user) — Small text files an agent can read (SVG, DXF, URDF, JSON, CSV; up to 32 KB) also come inline as an embedded resource; STEP, STL, 3MF and GLB are link-only — do not fetch them into your context, give the user the link. `output_path` only names the file; the returned `output_path` is a server-side path, usable only as `file` in later calls to this server. HEAVY FORMATS (step/stl/3mf/glb): if OCCT serialization outlives the client tool-call limit, this tool returns { ok:true, status:"running", code:"export.deferred", job_id } while the export continues on the server — poll by calling export({ job_id }) until status is done/failed. CRITICAL: open_in_studio / paint does NOT need STEP. On export.deferred or any STEP stall, call open_in_studio immediately for the shaded widget; do not block paint on export.

  ConnectorNo auth

• calculateA

  utility-converterA

  Safely evaluate an arithmetic expression and return the exact result. Supports +, -, *, /, //, %, ** (power), parentheses, the constants pi and e, and the functions sqrt(), abs(), round(). Does not execute arbitrary code.

  ConnectorNo auth

• goal-dismiss-red-teamA

  Planner (proof-of-done)A

  Dismiss the red-team finding that blocks a goal from starting, when you disagree with the verdict. Requires a reason of at least 80 characters explaining why the counterexample does not apply. The dismissal is recorded in goal history and surfaced in goal-get, so "agent overrode the gate" stays distinguishable from "no hole was found". Prefer fixing the acceptance criteria via goal-add-criterion — dismissing leaves the hole open.

  ConnectorNo auth

• research_regimeA

  f1owA

  Deterministic market-regime classification across seven axes: trend (Mayer Multiple bands vs the 200-day SMA), volatility (30d realized, percentile vs own history), cycle (a ~30-indicator top checklist incl. MVRV Z-Score, NUPL, Pi Cycle, AHR999), sentiment (Fear & Greed level, trajectory and percentile), leverage stress (derivatives-risk-index percentile), liquidity (7d spot-ETF net flows vs prior 7d), and rotation (Altcoin Season index, BTC dominance, ETH/BTC 30d). Each axis reports its state, the numbers behind it, and a +1/0/-1 vote; the overall posture is the disclosed sum-of-votes rule, with one exception — while 30d realized volatility sits at or below its 5th percentile the tape is compressed, trend structure stops being tradeable, and the Trend axis's vote is withheld (its reading is still reported). Compression is flagged structurally, and adds an informational, never-voting asymmetry lean naming which side is more exposed. Includes explicit 'would change the call' thresholds to watch. Rule-based and deterministic — same inputs, same read. Useful as a first call: most other signals (funding, sentiment, flows) read differently depending on this regime.

  ConnectorNo auth

• goal-dismiss-red-teamA

  Planner MCP — provable "done" for AI agentsA

  Dismiss the red-team finding that blocks a goal from starting, when you disagree with the verdict. Requires a reason of at least 80 characters explaining why the counterexample does not apply. The dismissal is recorded in goal history and surfaced in goal-get, so "agent overrode the gate" stays distinguishable from "no hole was found". Prefer fixing the acceptance criteria via goal-add-criterion — dismissing leaves the hole open.

  ConnectorNo auth

• calculateA

  stampA

  Evaluate a math expression safely: + - * / // % **, parentheses, functions (abs round min max sqrt floor ceil exp log log2 log10 pow sin cos tan), constants pi e tau inf. Parsed to an AST and only whitelisted nodes are computed -- no eval(), no names, no arbitrary code.

  ConnectorNo auth

• search_grantsA

  opennihA

  Use this tool when the user asks to find NIH grants — by topic, principal investigator, institution, NIH Institute/Center, activity code, fiscal year, or project number ('what NIH grants exist on CRISPR', 'show me Stanford's R01s in 2022', 'look up 5R01CA036569-02'). This is the primary entry point: results carry pi_profile_id, the NIH RePORTER PI identifier. Read-only bounded search over the entity-resolved NIH ExPORTER/RePORTER corpus; results include mechanism_class, funding_basis and provenance, and a null award_amount means not reported, never zero. Bounds: query/text_search/institution/pi_name 2-64 chars, project_num ^[A-Z0-9-]{1,24}$, activity_code ^[A-Z0-9]{1,5}$, fiscal years 1985-2100, limit 1-50 (default 20), offset 0-100000 (offset>10000 is rate-limited to 12/min — partition by fiscal_year instead). meta.total_funding is the full matching-slice sum, not the returned page. Results are cached for ~1 hour. Counting: one result is one source application-year row, so (project_num, fiscal_year) is NOT unique - a multi-component award (P/U center) emits one row per sub-project, so the same project_num can appear several times in one fiscal year with different PIs and titles. These are distinct NIH application records, not duplicates. Each row carries appl_id, the NIH application id and authoritative row identity - two rows sharing (project_num, fiscal_year) with different appl_id are distinct sub-project applications, and appl_id is the key to de-duplicate on. meta.total counts rows, not grants; meta.unique_project_nums counts distinct project numbers, and meta.distinct_awards (row field core_project_num) counts distinct awards.

  ConnectorNo auth

• calculateA

  teardropA

  Evaluate a safe arithmetic expression and return the numeric result. Supports +,-,*,/,**,%,sqrt,abs,round,floor,ceil,log,sin,cos,tan,pi,e. Use for derived figures (ratios, differences, percentages) that no tool output already provides. Use when: Use when an agent must derive a value that tool outputs do not already include, such as a ratio, spread, or percentage. Do not use to re-derive pre-computed fields like std_30d or dca_baseline_90d that get_token_price_historical returns. Limitations: Arithmetic only — no aggregation over tool history, no unit conversion, and no financial modeling beyond the listed functions. Errors return an error field rather than raising. Alternatives: get_token_price_historical, get_datetime

  ConnectorNo auth

• import_handoffA

  kernelcadA

  Use this when you need to place parts from a LabWired handoff.json into a KernelCAD assembly. Pass the parsed handoff object using schema labwired-kernelcad-handoff/0.1. The tool validates the handoff and returns an evaluable assembly script with corner-anchored bbox placeholder solids, origin/bbox-face/mounting-hole/pin anchors, endpoint connectivity mates for wires whose two endpoint positions are known, and structured cutout_hints from connector faces. These mates express connectivity intent only; no wire geometry or routing is generated. It does not cut enclosure walls or fetch STEP/GLB assets. position_mm locates the bbox lower corner; bbox.center is derived. Null, estimated, proposed, and other non-exact geometry is reported as warnings and omitted or placed at zero as appropriate.

  ConnectorNo auth