nexo-brain 1.4.1 → 1.5.1

package/README.md

@@ -198,7 +198,7 @@ This means long sessions (8+ hours) feel like one continuous conversation instea
 
 ## Cognitive Features
 
-NEXO Brain provides 29 cognitive tools on top of the 78 base tools, totaling **115+ MCP tools**. These features implement cognitive science concepts that go beyond basic memory:
+NEXO Brain provides **100+ MCP tools** implementing cognitive science concepts that go beyond basic memory:
 
 ### Input Pipeline
 
@@ -226,11 +226,13 @@ NEXO Brain provides 29 cognitive tools on top of the 78 base tools, totaling **1
 |---------|-------------|
 | **HyDE Query Expansion** | Generates hypothetical answer embeddings for richer semantic search. Instead of searching for "deploy error", it imagines what a helpful memory about deploy errors would look like, then searches for that. |
 | **Hybrid Search (FTS5+BM25+RRF)** | Combines dense vector search with BM25 keyword search via Reciprocal Rank Fusion. Outperforms pure semantic search on precise terminology and code identifiers. |
+| **KG Boost** | Knowledge Graph connection count influences retrieval ranking. Memories linked to well-connected entities (many edges) receive a logarithmic score bonus, surfacing contextually important facts higher. |
+| **HNSW Vector Index** | Optional approximate nearest neighbor index (hnswlib). Activates automatically when memory count exceeds 10,000. Falls back to exact brute-force below that threshold — no configuration needed. |
 | **Cross-Encoder Reranking** | After initial vector retrieval, a cross-encoder model rescores candidates for precision. The top-k results are reordered by true semantic relevance before being returned to the agent. |
 | **Multi-Query Decomposition** | Complex questions are automatically split into sub-queries. Each component is retrieved independently, then fused for a higher-quality answer — improves recall on multi-faceted prompts. |
 | **Temporal Indexing** | Memories are indexed by time in addition to semantics. Time-sensitive queries ("what did we decide last Tuesday?") use temporal proximity scoring alongside semantic similarity. |
 | **Spreading Activation** | Graph-based co-activation network. Memories retrieved together reinforce each other's connections, building an associative web that improves over time. |
-| **Recall Explanations** | Transparent score breakdown for every retrieval result. Shows exactly why a memory was returned: semantic similarity, recency, access frequency, and co-activation bonuses. |
+| **Recall Explanations** | Transparent score breakdown for every retrieval result. Shows exactly why a memory was returned: semantic similarity, recency, access frequency, KG boost, and co-activation bonuses. |
 
 ### Proactive
 
@@ -412,7 +414,7 @@ That's it. No need to run `claude` manually. Your operator will greet you immedi
 | Component | What | Where |
 |-----------|------|-------|
 | Cognitive engine | Python: fastembed, numpy, vector search | pip packages |
-| MCP server | 111+ tools for memory, cognition, learning, guard | ~/.nexo/ |
+| MCP server | 100+ tools for memory, cognition, learning, guard | ~/.nexo/ |
 | Plugins | Guard, episodic memory, cognitive memory, entities, preferences | ~/.nexo/plugins/ |
 | Hooks (6) | SessionStart briefing, Stop post-mortem, PostToolUse capture, PreCompact checkpoint, PostCompact recovery, Caffeinate | ~/.nexo/hooks/ |
 | Reflection engine | Processes session buffer, extracts patterns, updates user model | ~/.nexo/scripts/ |
@@ -425,12 +427,23 @@ That's it. No need to run `claude` manually. Your operator will greet you immedi
 
 - **macOS or Linux** (Windows via [WSL](https://learn.microsoft.com/en-us/windows/wsl/install))
 - **Node.js 18+** (for the installer)
-- **Claude Opus (latest version) strongly recommended.** NEXO Brain provides 111+ MCP tools across 20 categories. This cognitive load requires a top-tier model with large context window. Smaller models (Haiku, Sonnet) may struggle with tool selection and produce inconsistent results. Opus handles all 111+ tools without hesitation.
+- **Claude Opus (latest version) strongly recommended.** NEXO Brain provides 100+ MCP tools across 20 categories. This cognitive load requires a top-tier model with large context window. Smaller models (Haiku, Sonnet) may struggle with tool selection and produce inconsistent results. Opus handles all 100+ tools without hesitation.
 - Python 3, Homebrew, and Claude Code are installed automatically if missing.
 
 ## Architecture
 
-### 111+ MCP Tools across 20 Categories
+### Modular Package Structure (v1.5.0)
+
+The core is organized into two Python packages:
+
+| Package | Modules | Responsibility |
+|---------|---------|----------------|
+| `db/` | 11 modules (`_core`, `_schema`, `_sessions`, `_learnings`, `_episodic`, `_credentials`, `_entities`, `_evolution`, `_fts`, `_reminders`, `_tasks`) | All SQLite persistence: schema migrations, CRUD, FTS indexing |
+| `cognitive/` | 6 modules (`_core`, `_memory`, `_ingest`, `_search`, `_decay`, `_trust`) | Cognitive engine: embeddings, RAG, decay, trust scoring |
+
+The rest of the server (`server.py`, `tools_*.py`, `plugins/`) stays flat for clarity.
+
+### 100+ MCP Tools across 20 Categories
 
 | Category | Count | Tools | Purpose |
 |----------|-------|-------|---------|
@@ -439,7 +452,7 @@ That's it. No need to run `claude` manually. Your operator will greet you immedi
 | Cognitive Advanced | 8 | hyde_search, spread_activate, explain_recall, dream, prospect, hook_capture, pin, archive | Advanced retrieval, proactive, lifecycle |
 | Guard | 3 | check, stats, log_repetition | Metacognitive error prevention |
 | Episodic | 10 | change_log/search/commit, decision_log/outcome/search, review_queue, diary_write/read, recall | What happened and why |
-| Sessions | 4 | startup, heartbeat, stop, status | Session lifecycle + context shift detection |
+| Sessions | 4 | startup, heartbeat, stop, status | Session lifecycle + context shift detection + inter-terminal auto-inbox |
 | Coordination | 7 | track, untrack, files, send, ask, answer, check_answer | Multi-session file coordination + messaging |
 | Reminders | 5 | list, create, update, complete, delete | User's tasks and deadlines |
 | Followups | 4 | create, update, complete, delete | System's autonomous verification tasks |
@@ -455,6 +468,7 @@ That's it. No need to run `claude` manually. Your operator will greet you immedi
 | Adaptive & Somatic | 4 | adaptive_weights, adaptive_override, somatic_check, somatic_stats | Learned signal weights + pain memory per file |
 | Knowledge Graph | 4 | kg_query, kg_path, kg_neighbors, kg_stats | Bi-temporal entity-relationship graph |
 | Context Continuity | 2 | checkpoint_save, checkpoint_read | Auto-compaction session preservation |
+| Claim Graph | — | (internal) | Atomic facts with provenance and contradiction detection |
 
 ### Plugin System
 
@@ -512,7 +526,7 @@ NEXO Brain is designed as an MCP server. Claude Code is the primary supported cl
 npx nexo-brain
 ```
 
-All 111+ tools are available immediately after installation. The installer configures Claude Code's `~/.claude/settings.json` automatically.
+All 100+ tools are available immediately after installation. The installer configures Claude Code's `~/.claude/settings.json` automatically.
 
 ### OpenClaw
 
@@ -591,11 +605,11 @@ NEXO Brain builds on ideas from several open-source projects. We're grateful for
 
 | Project | Inspired Features |
 |---------|------------------|
-| [Vestige](https://github.com/pchaganti/gx-vestige) | HyDE query expansion, spreading activation, prediction error gating, memory dreaming, prospective memory |
-| [ShieldCortex](https://github.com/PShieldCortex/ShieldCortex) | Security pipeline (4-layer memory poisoning defense) |
-| [Bicameral](https://github.com/nicobailey/Bicameral) | Quarantine queue (trust promotion policy for new facts) |
-| [claude-mem](https://github.com/nicobailey/claude-mem) | Hook auto-capture (extracting decisions and facts from conversations) |
-| [ClawMem](https://github.com/nicobailey/ClawMem) | Co-activation reinforcement (memories retrieved together strengthen connections) |
+| Vestige | HyDE query expansion, spreading activation, prediction error gating, memory dreaming, prospective memory |
+| ShieldCortex | Security pipeline (4-layer memory poisoning defense) |
+| Bicameral | Quarantine queue (trust promotion policy for new facts) |
+| claude-mem | Hook auto-capture (extracting decisions and facts from conversations) |
+| ClawMem | Co-activation reinforcement (memories retrieved together strengthen connections) |
 
 ## Support the Project
 
@@ -610,6 +624,43 @@ If NEXO Brain is useful to you, consider:
 
 ## Changelog
 
+### v1.5.0 — Modular Core + Knowledge Graph Search (2026-03-29)
+- **Architecture**: `db.py` refactored into `db/` package (11 modules: core, schema, sessions, learnings, episodic, credentials, entities, evolution, fts, reminders, tasks)
+- **Architecture**: `cognitive.py` refactored into `cognitive/` package (6 modules: core, memory, ingest, search, decay, trust)
+- **KG Boost**: Knowledge Graph connection count now influences search result ranking — well-connected entities surface higher in retrieval
+- **HNSW Vector Index**: Optional approximate nearest neighbor acceleration (activates automatically above 10,000 memories, falls back to brute-force otherwise)
+- **Claim Graph**: Decomposes blob memories into atomic verifiable facts with provenance, confidence scores, and contradiction detection
+- **Inter-terminal Auto-inbox (D+)**: `nexo_startup` now accepts `claude_session_id` (Claude Code session UUID) — enables automatic inbox delivery between parallel terminals via PostToolUse hook + migration v13
+- **Tests**: 24 pytest tests across 3 suites (cognitive, knowledge graph, migrations)
+
+### v1.4.1 — Multi-AI Code Review (2026-03-29)
+- **Fix**: 3 bugs found by GPT-5.4 (Codex CLI) + Gemini 2.5 (Gemini CLI) reviewing full codebase
+  - `session_diaries` → `session_diary` table name (smart startup silently failed)
+  - Quarantine contradiction logic distinguished confirmation from opposition
+  - Knowledge Graph timezone import crash
+- **Security**: Memory sanitization prevents prompt injection via stored content
+- **Migration #13**: Normalizes legacy status values (PENDIENTE→PENDING) on upgrade
+
+### v1.4.0 — The Brain Dreams (2026-03-29)
+- **Major**: All 9 nightly scripts migrated from Python word-overlap to CLI wrapper pattern
+  - Postmortem consolidator: opus understands patterns by meaning, not word overlap
+  - Sleep system: detects real duplicates via semantic understanding
+  - Daily synthesis: opus prioritizes what matters for tomorrow
+  - Self-audit: interprets findings for root cause analysis
+  - Evolution: prompt reduced 95% (45K → 2.2K chars) — CLI investigates using tools
+- **Stop Hook v8**: Session-scoped tool counting (not day-wide), buffer fallback removed
+- **Guard**: Behavioral rules section surfaces most-violated rules at session start
+- **Followup hygiene**: Weekly cleanup script normalizes statuses, flags stale items
+- 8 missing core scripts + 1 plugin added to repository
+
+### v1.3.0 — Evolution System (2026-03-28)
+- **New**: Self-improvement cycle — NEXO proposes and applies improvements weekly
+- Dual-mode: auto (low-risk changes) and review (owner approval required)
+- Circuit breaker, snapshot/rollback, immutable file protection
+
+### v1.2.3 — AGPL-3.0 License (2026-03-27)
+- License changed from MIT to AGPL-3.0
+
 ### v1.2.1 — Stop Hook Hotfix (2026-03-27)
 - **Fix**: v1.2.0 deleted the flag on approve, causing infinite block loops if session didn't close immediately
 - **Fix**: Removed TTL on flag — it persists until SessionStart cleans it up next session

package/package.json

@@ -1,8 +1,8 @@
 {
   "name": "nexo-brain",
-  "version": "1.4.1",
+  "version": "1.5.1",
   "mcpName": "io.github.wazionapps/nexo",
-  "description": "NEXO — Cognitive co-operator for Claude Code. Atkinson-Shiffrin memory, semantic RAG, trust scoring, and metacognitive error prevention.",
+  "description": "NEXO — Cognitive co-operator for Claude Code. Atkinson-Shiffrin memory, semantic RAG, knowledge graph, HNSW vector indexing, trust scoring, and metacognitive error prevention.",
   "bin": {
     "nexo-brain": "./bin/nexo-brain.js"
   },
@@ -14,6 +14,9 @@
     "ai-assistant",
     "vector-search",
     "atkinson-shiffrin",
+    "knowledge-graph",
+    "hnsw",
+    "claim-graph",
     "openclaw",
     "openclaw-plugin"
   ],
@@ -30,6 +33,7 @@
     "bin/",
     "src/",
     "templates/",
-    "scripts/"
+    "scripts/",
+    "tests/"
   ]
 }

package/src/claim_graph.py

@@ -0,0 +1,323 @@
+"""NEXO Claim Graph — Atomic claims with provenance and contradiction detection.
+
+Decomposes blob memories into individual verifiable facts. Each claim has a
+source memory, confidence score, and can be linked to contradicting claims.
+
+Tables (created in cognitive.db alongside KG):
+- claims: atomic facts with provenance
+- claim_links: relationships between claims (supports, contradicts, refines)
+"""
+
+import json
+import sqlite3
+import numpy as np
+from datetime import datetime, timezone
+from typing import Optional
+
+
+def _get_db():
+    """Get cognitive.db connection."""
+    import cognitive
+    return cognitive._get_db()
+
+
+def _embed(text: str) -> np.ndarray:
+    import cognitive
+    return cognitive.embed(text)
+
+
+def _cosine_similarity(a, b) -> float:
+    import cognitive
+    return cognitive.cosine_similarity(a, b)
+
+
+def _array_to_blob(arr: np.ndarray) -> bytes:
+    return arr.astype(np.float32).tobytes()
+
+
+def _blob_to_array(blob: bytes) -> np.ndarray:
+    return np.frombuffer(blob, dtype=np.float32)
+
+
+def init_tables():
+    """Create claim graph tables if they don't exist."""
+    db = _get_db()
+    db.executescript("""
+        CREATE TABLE IF NOT EXISTS claims (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            text TEXT NOT NULL,
+            embedding BLOB,
+            source_type TEXT NOT NULL DEFAULT '',
+            source_id TEXT NOT NULL DEFAULT '',
+            source_memory_store TEXT DEFAULT '',
+            source_memory_id INTEGER DEFAULT 0,
+            confidence REAL DEFAULT 1.0,
+            verification_status TEXT DEFAULT 'unverified',
+            verified_at TEXT,
+            domain TEXT DEFAULT '',
+            created_at TEXT DEFAULT (datetime('now')),
+            updated_at TEXT DEFAULT (datetime('now'))
+        );
+
+        CREATE TABLE IF NOT EXISTS claim_links (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            source_claim_id INTEGER NOT NULL REFERENCES claims(id),
+            target_claim_id INTEGER NOT NULL REFERENCES claims(id),
+            relation TEXT NOT NULL,
+            confidence REAL DEFAULT 1.0,
+            created_at TEXT DEFAULT (datetime('now')),
+            UNIQUE(source_claim_id, target_claim_id, relation)
+        );
+
+        CREATE INDEX IF NOT EXISTS idx_claims_source ON claims(source_type, source_id);
+        CREATE INDEX IF NOT EXISTS idx_claims_domain ON claims(domain);
+        CREATE INDEX IF NOT EXISTS idx_claims_status ON claims(verification_status);
+        CREATE INDEX IF NOT EXISTS idx_claim_links_source ON claim_links(source_claim_id);
+        CREATE INDEX IF NOT EXISTS idx_claim_links_target ON claim_links(target_claim_id);
+    """)
+    db.commit()
+
+
+def add_claim(text: str, source_type: str = "", source_id: str = "",
+              source_memory_store: str = "", source_memory_id: int = 0,
+              confidence: float = 1.0, domain: str = "") -> dict:
+    """Add an atomic claim to the graph.
+
+    Returns the claim dict with id, or existing claim if duplicate detected.
+    """
+    db = _get_db()
+    init_tables()
+
+    # Embed the claim
+    vec = _embed(text)
+    blob = _array_to_blob(vec)
+
+    # Check for near-duplicate claims (similarity > 0.92)
+    existing = find_similar_claims(text, threshold=0.92, limit=1)
+    if existing:
+        # Update confidence if new source provides additional evidence
+        dup = existing[0]
+        new_conf = min(1.0, dup["confidence"] + 0.1)
+        db.execute("UPDATE claims SET confidence = ?, updated_at = datetime('now') WHERE id = ?",
+                   (new_conf, dup["id"]))
+        db.commit()
+        return {"id": dup["id"], "action": "merged", "confidence": new_conf}
+
+    cursor = db.execute(
+        """INSERT INTO claims (text, embedding, source_type, source_id,
+           source_memory_store, source_memory_id, confidence, domain)
+           VALUES (?, ?, ?, ?, ?, ?, ?, ?)""",
+        (text, blob, source_type, source_id, source_memory_store,
+         source_memory_id, confidence, domain)
+    )
+    db.commit()
+    return {"id": cursor.lastrowid, "action": "added", "confidence": confidence}
+
+
+def find_similar_claims(text: str, threshold: float = 0.8, limit: int = 10) -> list[dict]:
+    """Find claims similar to the given text using cosine similarity."""
+    db = _get_db()
+    init_tables()
+
+    query_vec = _embed(text)
+    if np.linalg.norm(query_vec) == 0:
+        return []
+
+    rows = db.execute("SELECT * FROM claims").fetchall()
+    results = []
+    for row in rows:
+        if not row["embedding"]:
+            continue
+        vec = _blob_to_array(row["embedding"])
+        score = _cosine_similarity(query_vec, vec)
+        if score >= threshold:
+            d = dict(row)
+            d.pop("embedding", None)
+            d["similarity"] = round(score, 4)
+            results.append(d)
+
+    results.sort(key=lambda x: x["similarity"], reverse=True)
+    return results[:limit]
+
+
+def link_claims(source_id: int, target_id: int, relation: str,
+                confidence: float = 1.0) -> dict:
+    """Create a relationship between two claims.
+
+    Relations:
+        - 'supports': source provides evidence for target
+        - 'contradicts': source contradicts target
+        - 'refines': source is a more specific version of target
+        - 'supersedes': source replaces target (target is outdated)
+    """
+    db = _get_db()
+    init_tables()
+
+    valid_relations = {"supports", "contradicts", "refines", "supersedes"}
+    if relation not in valid_relations:
+        return {"error": f"Invalid relation. Use: {valid_relations}"}
+
+    try:
+        db.execute(
+            "INSERT OR IGNORE INTO claim_links (source_claim_id, target_claim_id, relation, confidence) "
+            "VALUES (?, ?, ?, ?)",
+            (source_id, target_id, relation, confidence)
+        )
+        db.commit()
+
+        # If contradiction, update verification status
+        if relation == "contradicts":
+            db.execute(
+                "UPDATE claims SET verification_status = 'contradicted', updated_at = datetime('now') WHERE id = ?",
+                (target_id,)
+            )
+            db.commit()
+
+        return {"source": source_id, "target": target_id, "relation": relation}
+    except Exception as e:
+        return {"error": str(e)}
+
+
+def detect_contradictions(claim_id: int) -> list[dict]:
+    """Find claims that potentially contradict the given claim."""
+    db = _get_db()
+    init_tables()
+
+    claim = db.execute("SELECT * FROM claims WHERE id = ?", (claim_id,)).fetchone()
+    if not claim:
+        return []
+
+    claim_vec = _blob_to_array(claim["embedding"])
+
+    # Find similar claims from different sources
+    rows = db.execute(
+        "SELECT * FROM claims WHERE id != ? AND domain = ?",
+        (claim_id, claim["domain"])
+    ).fetchall()
+
+    contradictions = []
+    for row in rows:
+        if not row["embedding"]:
+            continue
+        vec = _blob_to_array(row["embedding"])
+        sim = _cosine_similarity(claim_vec, vec)
+        # High similarity but different source = potential contradiction
+        if 0.6 <= sim <= 0.9 and row["source_id"] != claim["source_id"]:
+            d = dict(row)
+            d.pop("embedding", None)
+            d["similarity"] = round(sim, 4)
+            contradictions.append(d)
+
+    return contradictions
+
+
+def verify_claim(claim_id: int, status: str = "confirmed") -> dict:
+    """Update verification status of a claim.
+
+    Status: 'confirmed', 'contradicted', 'outdated', 'unverified'
+    """
+    db = _get_db()
+    init_tables()
+
+    valid = {"confirmed", "contradicted", "outdated", "unverified"}
+    if status not in valid:
+        return {"error": f"Invalid status. Use: {valid}"}
+
+    db.execute(
+        "UPDATE claims SET verification_status = ?, verified_at = datetime('now'), "
+        "updated_at = datetime('now') WHERE id = ?",
+        (status, claim_id)
+    )
+    db.commit()
+    row = db.execute("SELECT * FROM claims WHERE id = ?", (claim_id,)).fetchone()
+    if row:
+        d = dict(row)
+        d.pop("embedding", None)
+        return d
+    return {"error": f"Claim {claim_id} not found"}
+
+
+def get_claim(claim_id: int) -> Optional[dict]:
+    """Get a single claim with its links."""
+    db = _get_db()
+    init_tables()
+
+    row = db.execute("SELECT * FROM claims WHERE id = ?", (claim_id,)).fetchone()
+    if not row:
+        return None
+
+    d = dict(row)
+    d.pop("embedding", None)
+
+    # Get links
+    outgoing = db.execute(
+        "SELECT cl.*, c.text as target_text FROM claim_links cl "
+        "JOIN claims c ON c.id = cl.target_claim_id "
+        "WHERE cl.source_claim_id = ?", (claim_id,)
+    ).fetchall()
+    incoming = db.execute(
+        "SELECT cl.*, c.text as source_text FROM claim_links cl "
+        "JOIN claims c ON c.id = cl.source_claim_id "
+        "WHERE cl.target_claim_id = ?", (claim_id,)
+    ).fetchall()
+
+    d["links_out"] = [dict(r) for r in outgoing]
+    d["links_in"] = [dict(r) for r in incoming]
+    return d
+
+
+def search_claims(query: str = "", domain: str = "", status: str = "",
+                  limit: int = 20) -> list[dict]:
+    """Search claims by text, domain, and/or status."""
+    db = _get_db()
+    init_tables()
+
+    if query:
+        return find_similar_claims(query, threshold=0.5, limit=limit)
+
+    conditions = []
+    params = []
+    if domain:
+        conditions.append("domain = ?")
+        params.append(domain)
+    if status:
+        conditions.append("verification_status = ?")
+        params.append(status)
+
+    where = " AND ".join(conditions) if conditions else "1=1"
+    rows = db.execute(
+        f"SELECT id, text, source_type, source_id, confidence, verification_status, "
+        f"domain, created_at FROM claims WHERE {where} ORDER BY created_at DESC LIMIT ?",
+        params + [limit]
+    ).fetchall()
+    return [dict(r) for r in rows]
+
+
+def stats() -> dict:
+    """Claim graph statistics."""
+    db = _get_db()
+    init_tables()
+
+    total = db.execute("SELECT COUNT(*) FROM claims").fetchone()[0]
+    by_status = {}
+    for row in db.execute(
+        "SELECT verification_status, COUNT(*) as cnt FROM claims GROUP BY verification_status"
+    ).fetchall():
+        by_status[row["verification_status"]] = row["cnt"]
+    by_domain = {}
+    for row in db.execute(
+        "SELECT domain, COUNT(*) as cnt FROM claims WHERE domain != '' GROUP BY domain ORDER BY cnt DESC LIMIT 10"
+    ).fetchall():
+        by_domain[row["domain"]] = row["cnt"]
+    links = db.execute("SELECT COUNT(*) FROM claim_links").fetchone()[0]
+    contradictions = db.execute(
+        "SELECT COUNT(*) FROM claim_links WHERE relation = 'contradicts'"
+    ).fetchone()[0]
+
+    return {
+        "total_claims": total,
+        "by_status": by_status,
+        "by_domain": by_domain,
+        "total_links": links,
+        "contradictions": contradictions,
+    }

package/src/cognitive/__init__.py

@@ -0,0 +1,62 @@
+"""NEXO Cognitive Engine — Modular vector memory with Atkinson-Shiffrin model.
+
+This package replaces the monolithic cognitive.py. All public functions and
+constants are re-exported here for full backwards compatibility:
+    import cognitive
+    cognitive.search("query")
+    cognitive.embed("text")
+"""
+
+# Core: DB, embedding, cosine, constants, tables, redaction
+from cognitive._core import (
+    COGNITIVE_DB, EMBEDDING_DIM, LAMBDA_STM, LAMBDA_LTM,
+    PE_GATE_REJECT, PE_GATE_REFINE, _gate_stats,
+    DISCRIMINATING_ENTITIES,
+    POSITIVE_SIGNALS, NEGATIVE_SIGNALS, URGENCY_SIGNALS,
+    _get_db, _init_tables, _migrate_lifecycle, _migrate_co_activation,
+    _auto_migrate_embeddings,
+    _get_model, _get_reranker, rerank_results,
+    embed, cosine_similarity, _array_to_blob, _blob_to_array,
+    extract_temporal_date, redact_secrets,
+)
+
+# Search
+from cognitive._search import (
+    search, bm25_search, hyde_expand_query,
+    record_co_activation,
+    _kg_boost_results, _apply_temporal_boost,
+    create_trigger, check_triggers, list_triggers, delete_trigger, rearm_trigger,
+    # Constants
+    CO_ACTIVATION_DECAY, CO_ACTIVATION_BOOST, CO_ACTIVATION_MIN_STRENGTH,
+)
+
+# Ingest
+from cognitive._ingest import (
+    ingest, ingest_session, ingest_to_ltm, ingest_sensory,
+    prediction_error_gate, get_gate_stats, detect_patterns,
+    process_quarantine, quarantine_list, quarantine_promote, quarantine_reject, quarantine_stats,
+    security_scan,
+)
+
+# Decay and maintenance
+from cognitive._decay import (
+    apply_decay, promote_stm_to_ltm, gc_stm, gc_test_memories,
+    gc_sensory, gc_ltm_dormant, dream_cycle,
+)
+
+# Trust and sentiment
+from cognitive._trust import (
+    get_trust_events, auto_detect_trust_events,
+    detect_sentiment, log_sentiment,
+    get_trust_score, adjust_trust, get_trust_history,
+    detect_dissonance, resolve_dissonance, check_correction_fatigue,
+)
+
+# Memory operations
+from cognitive._memory import (
+    format_results, get_metrics, check_repeat_errors, rehearse_by_content,
+    consolidate_semantic, get_siblings,
+    get_stats, set_lifecycle, auto_merge_duplicates,
+    somatic_accumulate, somatic_guard_decay, somatic_nightly_decay,
+    somatic_project_events, somatic_get_risk, somatic_top_risks,
+)