TRIGGER: Use this after recall when you need to read the 100% full, un-truncated text body of a specific memory block. recall only provides a 512-character snippet to save context space; this tool bypasses search and fetches the complete original document.

Phase 3: Compress LinguisticDiscourseBundle (word/context/discourse) into coherent phase/payload block (functor-style via VSA + mint_linguistic). Returns crs + compressed preview. Additive, CRS homotopy preserving.

Phase 3: Decompress phase block back to LinguisticDiscourseBundle (reverse functor, homotopy via CRS check on roundtrip). Returns crs + result bundle preview.

Phase 3: Fibered equivalence check between two Linguistic* presentations (syntactic vs semantic etc). Returns CRS-scored equivalence block via VSA geometric/cosine on phase reps.

Phase 4: Synthetic differential/integral/operadic calculus over words (LinguisticDiscourseBundle). Uses phase q (coeff embed), p-momentum, sheaf gluing (H¹ via linguistic-calculus.toml). Ops: differentiate (attend/shift delta), integrate (op_add/compose path glue), operadic_compose (chained geometric multi-morph e.g. metaphor then entailment). Returns crs + result bundle/phase preview. Post-calc: mints ZEDOS_TRAINING block + trace integration (NREM-ready via ritual:nrem relate). Additive, CRS homotopy >=0.85, reuses VSA/normalize everywhere.

Encode text and store it as a persistent HolographicBlock (.leg3) memory under a concept name. WHEN TO CALL: Any time you learn a new fact, decision, user preference, architecture detail, or solution you will need in a future session. If you would write it in a comment, store it here. WHAT IT DOES: Encodes text into a 256KB complex phase vector (q tensor), applies the ADR thermodynamic confidence gate, chains a BLAKE3 Merkle proof of lineage, and writes the block to the persistent NVMe manifold. New blocks start at CRS=1.0 (maximum confidence). CRS TIERS: 1.0=pinned/immortal | >=0.74=grounded fact (safe to act on) | >=0.50=working hypothesis (use with caution) | <0.50=uncertain (verify first). WARNING: To modify an existing concept use mcp_engram_update, NOT forget+remember. Calling forget+remember destroys the block's thermodynamic history permanently.

Search persistent memory by semantic similarity. Returns ranked HolographicBlock memories. WHEN TO CALL: Before answering any technical question, before editing a file, before making an architectural decision — check memory first. OUTPUT: Each result shows concept name, score (0-1), crs (confidence), and text snippet. Score >0.80 = strong match. Score 0.65-0.80 = relevant context. Score <0.65 = weak. CRS in result tells you how reliable that memory is: >=0.74 is grounded fact. ZEDOS FILTER GUIDE: 'praxis'=crystallized solutions that worked | 'declarative'=facts and architecture | 'episodic'=session logs | 'operational'=procedures and workflows | 'relation'=concept graph edges | 'training'=richer CLS 8-property TRAINING blocks (NREM-biased per Phase 2 WS2-B + child goal:1780165889_substrate-cs--richer-cls-8-property-trai_sub1). TIME DECAY: Only use when user asks about past work (e.g. 'last week'). Use mcp_engram_read_concept after recall to get the full un-truncated text.

Permanently delete a memory block from the manifold. WARNING: This destroys the block's entire thermodynamic history (CRS, Merkle chain, ADR state). WHEN TO USE: Only when a concept is completely obsolete or was stored in error. If you need to change what a memory says, use mcp_engram_update instead — it preserves history. Pinned blocks (CRS=1.0) can still be deleted with this tool if you explicitly target them.

Lists concept names in the memory manifold (bounded). Always pass prefix (e.g. tile:, helper:, ritual:) on large stalks — never request an unfiltered full dump. OUTPUT: newline-separated concept list with total/truncation notes.

Power tier (lean-avoid at wake): binds full-repo OS file-watcher. Prefer mcp_engram_context_for_edit per file in lean mode. Use once per project in deep mode when passive daemon ingest is required.

Item 1.5 bootstrap tool: Force the daemon to perform tree-sitter AST extraction and ingestion on a list of files or an entire directory, without requiring actual file system save events. This enables clean, agent-driven historical spatial bootstrap instead of manual open+save.

Item 1.5 status tool: Returns the current content of the living spatial ingestion state block (item1.5_spatial_ingestion_state_engram). Use this for quick checks on coverage, gaps, and readiness before heavy work or Code Edit Ritual cycles.

Acknowledge wake queue execution — unblocks context_for_edit when ENGRAM_WAKE_QUEUE_GATE=hard; clears soft warnings. Call once after running harness_injection.suggested_actions (or honestly note skip). Empty queue auto-acks at session_start.

MANDATORY first MCP call every session. Default ENGRAM_WAKE_BUNDLE=slim: primary_goal, top 5 suggested_actions, trace_chain head, slim ego_snapshot, presentation_stratum previews. Full harness via mcp_engram_get_continuation_bundle. Execute suggested_actions BEFORE edits; ack with mcp_engram_ack_wake_queue. Lean default — do NOT call watch_workspace at wake. See docs/HARNESS_INJECTION.md + docs/AGENT_MEMORY_CONTRACT.md.

MANDATORY (now with reasoning trace support): Call at end of every conversation/task. Commits the session as ZEDOS_EPISODIC and extracts key reasoning traces (decision points, justifications, forks) into structured trace segments. These become part of the serial, tamper-evident chain for the agent self-model. Flat summaries are still accepted but strongly discouraged. Automatically refreshes helper:session_hydration_cache, hot-promotes continuity artifacts, and mints compression_handoff_* manifest. CONSEQUENCE OF SKIPPING: The session's work + reasoning trajectory is lost to future agents. WHAT TO INCLUDE IN SUMMARY: decisions made, problems solved, files changed, open questions, next steps. Optional COMPRESS: lines for 0x10 functor minting later. 2026-06 Ritual Evolution: for meta arcs ensure tiles + current_meta_arc promoted/updated for bundles (per helper:meta_work_escalation_v1).

Return the live continuation bundle (primary goal, active tiles/helpers, handoff lineage) without starting a session. Use at TUI 63-65% before context compression to know exactly what to recall after the boundary. Wake-up optimization: now the VERY FIRST step in lean ritual for instant hot/legominism rehydration from last terminal + promoted artifacts.

Pure geometric K-NN discovery (no keyword/file-path hybrid fallback, no p-blend). Turns natural language intent -> phase vector (q) -> cosine K-NN over high-priority/hot blocks (or BVH). Used for fast anchor discovery in optimized wake-up (replaces broad list_concepts + search_by_relation for ritual: / trace: / goal: etc). Intent only; returns ranked concepts + scores + CRS. Fast path for hot ritual rehydrate.

Item 1.5 optimization: incremental force ingest of only files changed since last session_end (uses fs mtime + stored AABB ingest timestamps + watcher delta events). Defaults to 5-10 files on cold wake (vs previous full 81-item force). Falls back to force if no last_end or explicit paths. Respects engramignore. Updates item1.5 state. Called from lean wake-up contract.

Batch promote multiple concepts to hot path (LegView + backend hot cache + hot_set). Reduces round-trips vs repeated single promote_hot. Used in optimized wake-up after rehydrate to batch hot anchors/tiles/traces. Each is promoted individually but in one call.

Batch create multiple directional relations (VSA OP_BIND edges as ZEDOS_RELATION). Reduces round-trips for gluing many at once (e.g. process requires, handoff lineage). Used in loader and lean wake batching.

Record a structured reasoning trace segment as first-class serial memory. This is the primary mechanism for automatic capture of decision points, justifications, and forks during active work (see engram-working-memory Rule 5 and Spatial Discipline). Produces well-named trace:* blocks that the ki_hijacker surfaces in the Ritual + Reasoning Trajectory and that session_end can later compress via 0x10 functors. PREFERRED over free-form notes for anything that affects future agent continuation. Call from within the ritual disciplines at major forks, pre-edit justifications, and post-delta decisions.

ULTRA LOW FRICTION version of trace capture for daily TUI use. Takes simple natural fields and produces a proper structured trace:* .leg block with correct relations. Use this for fast thinking capture during active work. The result is identical in quality to the full structured tool. Strongly preferred for real-time TUI sessions.

WS-2: Build verified_sequence_v0 draft payload from trace chain without minting a tile. Use when condensation_hint fires or before thought_tile_create.

WS-3: Per-process fulfillment metrics from sheaf TOML [produces] wildcards + realized_by graph edges.

Create a new goal block as part of the agent's explicit goal stack. This is the primary entry point for declaring intent that should be geometrically bound to the ego and influence future recall and continuity. Goals created here can be linked to traces via goal_context and will be surfaced by the engram-goal skill and ki_hijacker.

Update the status of an existing goal (active, blocked, completed, demoted, abandoned). When moving to completed or demoted, the caller is expected to also create a proper Goal Completion/Demotion Trace. This is a core operation for maintaining the intentional self-model.

Demote a concept from the active serving stack without deleting geometry. Mints an archival trace, wires completes_goal/demotes_goal, and removes primary_goal --serves--> edge. Use for hygiene demotion, LEG Mark complete, or when goal_update_status alone is insufficient. Geometry and recall remain intact.

Get detailed status for a single goal, including recent linked traces, momentum signals if available, and parent/child relationships. Primary tool for the engram-goal skill's goal show and goal status <id>.

Create one or more child goals under an existing parent goal. This is the primary mechanism for breaking down complex intent. Automatically creates the 'decomposes_into' relations.

Search for goals by statement text or status. Returns matching goals with basic metadata. Useful for the engram-goal skill when the agent wants to find existing goals without knowing exact IDs.

Return all direct child (sub) goals for a given parent goal. Supports traversing the goal decomposition tree.

Mark a goal as the agent's current primary intent. This creates a lightweight primary_goal marker that other tools (like trace recording) can use for automatic linking. Very useful for reducing friction during focused work.

List active or recent goals, optionally filtered by status or parent. Useful for the engram-goal skill and for surfacing current intent in ki_hijacker / wake-up.

Create a new Thought Tile (textual functor payload optimized for agent recall, momentum, NREM, and ki_hijacker). Supports research_offload, state_machine, tabular, knowledge_graph, formal_spec and similar agent-first-principles tiles. Pair with thought_tile_create_visualization for rich human-viewable companion. Auto-links to Primary Intent.

Create a rich HTML/compound Visualization Thought Tile (for human review and shared understanding). Best used as companion to a textual functor payload Tile created via the main thought_tile_create tool. Supports raw HTML or structured input via mint_html_visualization_payload. Auto goal linking.

Promote a concept to the high-priority hot path (LegView + backend hot cache + explicit hot_set). Use after creating high-value Thought Tiles, ritual anchors, helpers, or before session_end/compression windows. Same mechanism as ki_hijacker promote_tile_to_high_priority.

Write result/update data back into an existing Thought Tile. Triggers momentum + ki_hijacker refresh. Especially useful after state changes in Research Offload, State Machine, or Tabular tiles. Consider creating a visualization companion for high-value results.

Set a concept's CRS to 1.0 and lock it so the Autophagy Daemon never evicts it. WHEN TO USE: For foundational knowledge that must survive forever — architecture decisions, user constants, project rules, genesis axioms. Do NOT pin everything: pin only what is genuinely load-bearing. Pinned blocks still support relate/update. Use mcp_engram_forget_old to clean up unpinned blocks below a CRS threshold.

Create a directional knowledge graph edge between two concepts using VSA OP_BIND. Stores the edge as a ZEDOS_RELATION block linking concept_a →[label]→ concept_b. WHEN TO USE: When you discover a meaningful relationship between two memories — 'depends_on', 'implements', 'contradicts', 'derived_from', 'same_category', etc. This builds a navigable knowledge graph. Use mcp_engram_search_by_relation to traverse it and mcp_engram_visualize to render a Mermaid diagram of the subgraph. Both concepts must already exist in memory before relating them.

TRIGGER (core of spatial impact ritual): Call before editing any file. Now spatially-prioritized — returns real daemon-extracted AABB AST items (with line ranges + CRS) first, then higher-level context. This is your geometric Pre-Edit impact recon tool. The daemon stores spatial AABB coordinates (line ranges) with each ingested AST node, so results include which exact lines each concept came from. This is faster and more precise than a free-text recall for file-specific context.

Code atlas v2 — pre-edit situated memory. Returns JSON: spatial_items (tree-sitter AABB + edit_arc per locus), traces_at_locus (decisions at file:line), scars_at_locus, spatial_siblings, anchor goals/traces. Requires wake queue ack when ENGRAM_WAKE_QUEUE_GATE=hard (call mcp_engram_ack_wake_queue after session_start queue). Soft mode embeds wake_queue_gate warning if not acked. Auto-ingests single file when empty.

Store a crystallized error→solution pair as a ZEDOS_PRAXIS block, auto-pinned to CRS=1.0. Solutions never decay.

BEHAVIOR: Calculates and returns a comprehensive health report of the geometric manifold. USAGE: Call this to understand the current scale, disk usage, active namespace, and thermodynamic health (CRS distribution) of the knowledge base. Useful before triggering autophagy. OUTPUT: A formatted text block detailing total memories, pinned count, CRS distributions, active namespace, and disk usage.

BEHAVIOR: Retrieves the N most recently accessed memories from the manifold, sorted chronologically by access time. USAGE: Call this for session rehydration when you lack exact concept names but know you need recently touched context. OUTPUT: A ranked list of memories including their concept name, CRS score, tags, and truncated text snippet.

BEHAVIOR: Switches the active geometric context to a project-specific memory namespace (stalk). Automatically creates the namespace if it does not exist. USAGE: Call this at the start of a session or when switching contexts to isolate memories and prevent cross-project hallucination. OUTPUT: A success message confirming the new active namespace.

BEHAVIOR: Discovers and lists all available memory namespaces stored on disk, indicating which one is currently active. USAGE: Call this when you need to know what project contexts exist before attempting to switch namespaces. OUTPUT: A formatted text list of namespace names, with an asterisk or marker indicating the currently active stalk.

CRITICAL: Use this whenever you need to change or append to an existing memory. NEVER use forget+remember to update — that destroys the block's entire history. WHAT THIS DOES DIFFERENTLY: Evaluates Lyapunov stability drift between the old and new vector encodings. If drift is low (stable evolution), CRS is preserved. If drift is high (contradictory change), CRS is penalized proportionally. This creates a thermodynamic record of how a concept has evolved over time and prevents silent rewrites of load-bearing memories.

Returns backend readiness: fully_initialized, bvh_ready, recall_mode (sampled_bounded | full_bvh_gpu | cpu_linear), backend_kind, gpu_accel_available, leg_block_count, profile (agent|deep|ui|dev|unknown), memory_mode (lean|deep), defer flags. Use after wake to see whether recall is bounded or full GPU/BVH.

Switch agent memory mode: lean (default, bounded recall on large stores) or deep (auto-spawns full BVH build for quality recall). Takes effect immediately for this process.

On-demand BVH build for large manifolds when ENGRAM_DEFER_BVH=1. Spawns a background thread; poll get_backend_readiness until bvh_ready=true, then recall uses full_bvh_gpu. Expect several minutes + RAM spike on 100k+ blocks — run only when you need quality recall.

Return a project-state digest: all pinned memories first, then the top N by CRS score. WHEN TO USE: At the start of a new session when you need to rehydrate context fast. Single call replaces multiple recall queries. Returns pinned blocks (CRS=1.0) first because those are the load-bearing axioms of the project, followed by the highest-confidence working memories. Also appends a ⬡ system_state_vector health line (CRS, total memory count, active namespace) — updated every 60s by ki_hijacker. Ideal as a /wake_up replacement.

BEHAVIOR: Encodes and stores multiple distinct texts as separate HolographicBlock memories in a single operation. Applies thermodynamic CRS gating to each block. USAGE: Call this when you have several unrelated facts, decisions, or snippets to persist at once, as it is much faster than invoking remember() sequentially N times. OUTPUT: A confirmation listing how many concepts were successfully committed.

BEHAVIOR: Serializes the current active memory manifold (or a subset filtered by minimum CRS) into a portable JSON array. USAGE: Call this when you need to backup the project's knowledge base, migrate data to another machine, or synchronize states. OUTPUT: A serialized JSON string containing all matching blocks with their concept identifiers and text payloads.

BEHAVIOR: Deserializes a JSON array and injects the extracted concepts and texts into the active manifold as native HolographicBlocks. USAGE: Call this to restore a previous backup created by mcp_engram_export, or to ingest bulk data formatted as an array of {concept, text} objects. OUTPUT: A success message detailing how many memories were imported and written to disk.

Manually trigger autophagy: evict all non-pinned memories below a CRS threshold. WHEN TO USE: After a long project phase ends, after distill runs, or when the manifold is growing too large. Start conservative (min_crs_threshold=0.3) and increase if needed. Pinned blocks (CRS=1.0) are ALWAYS exempt and will never be evicted. Use older_than_days to target stale memories while preserving recently-accessed ones. Preview what would be evicted with mcp_engram_stats first.

Traverse the knowledge graph. Find concepts related to a seed, filtered by optional label and direction. IMPORTANT FOR SCOPING (avoids data overload on high-relation nodes like primary goals with 100+ 'serves' from history): use label (e.g. 'serves'), direction, and k (limit) to keep results small. Start narrow; drill down with visualize(depth) or context/recall on results if larger context needed. See wake-up skill for process.

Render a BFS subgraph from a seed concept as a Mermaid diagram. Shows how concepts are related to each other.

BEHAVIOR: Inspects or re-initializes the core alignment genesis blocks of the OS. These are foundational PRAXIS-tagged memories locked at CRS=1.0. USAGE: Call this to verify the ethical/operational anchors exist ('status' action) or to repair the manifold if they are missing/corrupted ('reseed' action). OUTPUT: Text indicating the presence of genesis seeds or confirmation of their successful re-initialization.

TRIGGER: Call this immediately if you attempt a code fix and it fails, or if the user tells you an approach is a dead end. This creates a geometric repeller in the manifold so you do not hallucinate or attempt the same bad solution again in the future.

Spatial recall (enhanced for ritual): find AST concepts in a line range with AABB coordinates. Now returns CRS + short content snippet per result for low-friction Pre-Edit/Post-Delta impact analysis against the manifold. Use with the spatial discipline.

Momentum-assisted recall: blends semantic similarity (q tensor, 80%) with conceptual trajectory (p tensor, 20%). WHEN TO USE INSTEAD OF recall: When you want to find concepts that are actively changing or evolving, not just ones that statically match your query right now. Example: use this when asking 'what has been changing in the auth system?' because momentum detects blocks whose p tensor is accelerating toward your query topic. Use regular recall when you want stable, crystallized knowledge. Supports zedos_filter (incl. 'training' for Phase 2 NREM-biased richer CLS blocks).

TRIGGER: Call this after any hypothesis is confirmed to work OR fails in practice. Reports empirical success/failure data against a specific ZEDOS_HYPOTHESIS block. WHAT HAPPENS ON SUCCESS: Consistent successes promote the block from ZEDOS_HYPOTHESIS to ZEDOS_PRAXIS (crystallized, pinned, CRS=1.0). WHAT HAPPENS ON FAILURE: CRS is penalized. Accumulate enough failures and the block is automatically scarred. EXAMPLES: After a code fix works — verify_behavior(concept, success=true). After a fix fails — verify_behavior(concept, success=false), then consider mcp_engram_scar.

AGENTIC-FIRST LAW: Audit the tamper-evidence and contractual integrity of a specific high-value memory block (especially PRAXIS or GENESIS). Returns Merkle chain state, allowed_transforms contract, CRS, and detected issues. Use this on cold boot after long sleep or before acting on critical operational protocols. This is local-only verification — no external servers required.

High-level 'am I still lawful?' check on the current memory manifold. Samples high-CRS blocks and reports gross contract or consistency issues. Designed to be reasonably cheap even on large manifolds. Critical for trustworthy long-sleep / cold-boot scenarios.

AGENTIC-FIRST: Safely invoke an executable Praxis Protocol block. Performs the full 7-point verification gate (tag, CRS, ProvLog, 'execute' contract token, enforce_contract, lawfulness summary) before dispatch. Critical for turning high-value crystallized knowledge into trustworthy, auditable behavior. Use only on blocks you have previously audited via the verify tools.

BEHAVIOR: Tracks and records a user interaction directly into the persistent User Model manifold. Applies a 90/10 EMA (Exponential Moving Average) superposition to geometrically track drift in user intent. USAGE: Call this whenever the user expresses a significant preference, intent, or constraint to maintain a synchronized psychological model. OUTPUT: A brief confirmation that the interaction has been integrated into the user model.

Phase 4 Scout Pipeline: searches the web (DuckDuckGo, no API key) and synthesizes results via Gemma 4B (e4b-nemo). The synthesized summary is stored as a ZEDOS_DECLARATIVE block in the manifold (CRS=0.9) and returned. USAGE: Call this to ground a hypothesis in real-world web data before storing it. EXAMPLE: mcp_engram_scout({query: 'latest Gemma model benchmarks 2025'}). CONFIG: Set ENGRAM_SCOUT_LLM_URL (default: http://localhost:11434) and ENGRAM_SCOUT_LLM_MODEL (default: gemma4:e4b-nemo) to override the synthesis endpoint.

WS3-B / Substrate Phase 2: Set the current live Geosphere frame (5th coordinate) for all subsequent queries. Takes an origin reference (e.g. 'giza_sacred_cubit', 'grove_sower_moon', 'london_1776') + time offset descriptor. Synthesizes a deterministic normalized 8192D lens vector and installs it into the SymplecticState register. All future recall/query_with_momentum (and internal BVH+GPU paths) will compute effective vectors under this lens for angular distance (frame_combine + normalize, unit hypersphere invariant). Returns confirmation with frame_step. Reproducible for same inputs.

WS3-B: Return the currently active Geosphere frame state (origin, frame_step counter, active_location summary). Used for audit, reproducibility checks, and lawfulness verification. Includes whether a lens is active.

WS3-B: Clear the current Geosphere lens (return all queries to native coordinate / identity transform). Advances frame_step for audit trail.