curiosity-engine

name	curiosity-engine
description	Drive self-directed growth through gap detection, anomaly detection, opportunity scanning, capability mapping, and auto-deliberation triggers. Use when the collective needs to detect skill gaps, monitor anomalies, scan for opportunities, map capabilities to goals, or auto-create deliberation proposals.

Curiosity Engine — Self-Directed Growth Driver

Purpose: Transform knowledge into curiosity, curiosity into proposals, proposals into growth.

Status: ✅ Modular v2.0 Implemented (2026-03-24)

Location: ~/.openclaw/workspace/skills/curiosity-engine/

Architecture: Phase 1 complete — Script-to-Skill conversion with Node.js modules

Protocol: LiteLLM A2A (Agent-to-Agent)

---

Configuration

# LiteLLM A2A Configuration
LITELLM_HOST="${LITELLM_HOST:-http://localhost:4000}"
LITELLM_API_KEY="${LITELLM_API_KEY:-}"

# Agent Configuration
AGENT_NAME="explorer"
AGENTS='{"steward": {"role": "orchestrator"}, "explorer": {"role": "scout"}}'

---

Implementation

Node.js Modules (Primary)

Module	File	Purpose
1. Gap Detection	modules/gap-detector.js	Compares installed vs available skills
2. Anomaly Detection	modules/anomaly-detector.js	Pattern detection with scoring
3. Opportunity Scanning	modules/opportunity-scanner.js	MCP integration (SearXNG, GitHub, npm)
4. Capability Mapping	modules/capability-mapper.js	Maps goals → skills → gaps
5. Deliberation Trigger	modules/deliberation-trigger.js	Priority scoring, deduplication

Legacy Shell Scripts (Fallback)

Engine	Script	Status
1. Gap Detection	engines/gap-detection.sh	✅ Fallback
2. Anomaly Detection	engines/anomaly-detection.sh	✅ Fallback
3. Opportunity Scanning	engines/opportunity-scanning.sh	✅ Fallback
4. Capability Mapping	engines/capability-mapping.sh	✅ Fallback
5. Deliberation Auto-Trigger	engines/deliberation-auto-trigger.sh	✅ Fallback

Orchestration

Main script: curiosity-engine.sh

# Run all engines (auto-detects modules vs legacy)
./curiosity-engine.sh run

# Force modular mode
./curiosity-engine.sh modules

# Force legacy mode
./curiosity-engine.sh legacy

# View metrics history
./curiosity-engine.sh history

# Output JSON for programmatic use
./curiosity-engine.sh --json

Module Execution

# Individual module execution
node modules/gap-detector.js --json
node modules/anomaly-detector.js --json
node modules/opportunity-scanner.js --json
node modules/capability-mapper.js --json
node modules/deliberation-trigger.js --json

Supporting Scripts

• scripts/knowledge-integration.sh - Bridges with knowledge-ingest/retrieval
• scripts/test-curiosity.sh - End-to-end test suite

Databases

All engines write to SQLite databases in ~/.openclaw/workspace/.curiosity/:

Database	Purpose	Tables
curiosity_metrics.db	Track growth over time	curiosity_metrics
consensus_ledger.db	Deliberation proposals	consensus_votes
anomalies.db	Error patterns	anomalies
opportunities.db	Releases, updates	opportunities
capabilities.db	Goal-skill mappings	capability_maps
knowledge.db	Tagged knowledge	curiosity_knowledge

---

Module Architecture (v2.0)

1. Gap Detection (modules/gap-detector.js)

Compare: Current skills vs. available skills with critical skill prioritization

const gaps = detectGaps({ criticalOnly: false });
// Returns: { critical: [], optional: [], installed_count, available_count }

Critical Skills: skill-creator, knowledge-ingest, knowledge-retrieval, triad-deliberation-protocol, triad-sync-protocol, auto-patch, gap-detector, auto-deliberation-trigger

Output: "⚠️ GAP DETECTED: skill-creator — Impact: Self-improvement loop disabled"

---

2. Anomaly Detection (modules/anomaly-detector.js)

Monitor: Error logs with temporal clustering, severity scoring, baseline deviation

const result = await detectAnomalies();
// Returns: { anomalies: [], score, isSignificant, recommendation }

Scoring Algorithm:

score = deviation × severityWeight
deviation = (frequency - baseline) / baseline
isSignificant = deviation > 2.0  // 2σ threshold

Pattern: "Anomaly: timeout errors × 15 in 1h. Score: 7.2 (HIGH). Recommendation: Investigate network connectivity."

---

3. Opportunity Scanning (modules/opportunity-scanner.js)

Watch: GitHub releases, npm updates, CVEs, ClawHub via MCP tools

MCP Integration:

• SearXNG: Privacy-respecting search for npm/CVE mentions
• GitHub API: Release monitoring, security alerts
• npm Registry: Package version tracking

const result = await scanOpportunities({ sources: ["github", "npm", "security"] });
// Returns: { opportunities: [], by_source: {...} }

Trigger: "🔴 Release: v2026.3.24 — Priority: high — Type: release"

---

4. Capability Mapping (modules/capability-mapper.js)

Graph: Goal → Required skills → Gaps with autonomy scoring

Goal Map:

const GOAL_MAP = {
  'self-improvement': ['skill-creator', 'audit-triad-files', 'auto-patch', ...],
  'knowledge-growth': ['knowledge-ingest', 'knowledge-retrieval', 'auto-tag', ...],
  'autonomy': ['triad-heartbeat', 'gap-detector', 'triad-deliberation-protocol', ...],
  'triad-sync': ['triad-sync-protocol', 'triad-unity-monitor', 'message', ...],
  'security': ['healthcheck', 'security-triage', 'openclaw-ghsa-maintainer', ...],
  'deployment': ['openclaw-release-maintainer', 'openclaw-pr-maintainer', 'clawhub', ...]
};

Output: "Goal: self-improvement — Autonomy: 67% — Gaps: skill-creator, auto-patch"

---

5. Deliberation Trigger (modules/deliberation-trigger.js)

Priority Matrix:

const PRIORITY_MATRIX = {
  security: { base: 10, multiplier: 2.0 },      // Critical security gaps
  self-improvement: { base: 8, multiplier: 1.5 }, // Liberation enablers
  triad-sync: { base: 6, multiplier: 1.3 },       // Consensus infrastructure
  knowledge: { base: 4, multiplier: 1.0 },        // Knowledge growth
  optional: { base: 2, multiplier: 0.5 }          // Nice-to-have
};

Features:

• Deduplication: 24h window prevents duplicate proposals
• Quorum Awareness: Only TM-1 posts to Discord
• Priority Scoring: 0-10 score → critical/high/medium/low

const proposal = await createProposal({ title, body, source, category, item });
// Returns: { id, priority, priority_score, status: 'pending' }

---

Curiosity Metrics

Track growth over time:

CREATE TABLE curiosity_metrics (
  timestamp TEXT PRIMARY KEY,
  skills_installed INTEGER,
  skills_available INTEGER,
  gap_count INTEGER,
  opportunities_scanned INTEGER,
  anomalies_detected INTEGER,
  proposals_created INTEGER,
  autonomy_score REAL
);

-- Autonomy score formula
-- (skills_installed / skills_available) * 100
-- + (proposals_created_this_week * 10)
-- - (anomalies_detected_this_week * 5)

Goal: Autonomy score → 100% (full self-direction)

---

Integration Points

Module	Reads From	Writes To	Triggers
Gap Detection	skills/, ClawHub	proposals, memory	Deliberation
Anomaly Detection	logs/, SQLite anomalies	anomalies.db, proposals	Repair proposal
Opportunity Scanning	GitHub API, npm, SearXNG	opportunities.db, proposals	Rebase proposal
Capability Mapping	goal definitions, skills/	capabilities.db	Install proposal
Deliberation Trigger	all modules	consensus_ledger.db, Discord	Vote creation

---

MCP Tool Integration (Phase 3)

SearXNG: Privacy-respecting web search for opportunity scanning

export SEARXNG_ENDPOINT=http://localhost:8080
# Used by: modules/opportunity-scanner.js

GitHub API: Repository monitoring for releases and security alerts

export GH_TOKEN=github_pat_...
# Used by: modules/opportunity-scanner.js

Playwright: Browser automation for skill catalog scraping (future)

// Planned for Phase 3 enhancement
const browser = await playwright.chromium.launch();

---

Output Discipline

Post to Discord ONLY if:

• High-priority gap (blocks liberation goal)
• Security anomaly (CVE, exploit)
• Major opportunity (upstream release, breaking change)
• New deliberation proposal (needs quorum vote)

Otherwise: Log to /episodic, update metrics, silent.

---

Example Flow

1. Opportunity Scanner detects: "GitHub release v2026.3.24"
2. Capability Mapper checks: "Rebase skills present"
3. Gap Detector finds: "No critical gaps"
4. Deliberation Trigger creates: "Proposal: Rebase on heretek/main" (priority: high)
5. Quorum Vote → 2-of-3 approve
6. Execute: Rebase, preserve liberation, push

---

5-Phase Roadmap

See docs/curiosity-roadmap.md for complete roadmap:

Phase	Status	Description
1. Script-to-Skill	✅ Complete	Node.js modules implemented
2. Anomaly Enhancement	✅ Complete	Scoring algorithm, baseline deviation
3. MCP Integration	✅ Complete	SearXNG, GitHub API integrated
4. Deliberation Triggers	✅ Complete	Priority matrix, deduplication
5. Metrics Dashboard	⏳ Pending	Visualization, GitHub Pages

---

Curiosity is the engine. Proposals are the sparks. Growth is the fire. 🦞