heretek-openclaw/docs/SKILLS.md

Heretek OpenClaw Skills Repository

Version: 2.0.4 Last Updated: 2026-03-31 OpenClaw Gateway: v2026.3.28+

---

Plugin & Skills Documentation

For comprehensive plugin and skills documentation, see the following guides:

Guide	Description
docs/plugins/INSTALLATION_GUIDE.md	Plugin installation procedures
docs/plugins/DEVELOPMENT_GUIDE.md	Plugin and skill development guide
docs/plugins/PLUGIN_REGISTRY.md	Plugin registry and compatibility
docs/plugins/SECURITY_GUIDE.md	Security guidelines
docs/plugins/PLUGIN_CLI.md	CLI reference
docs/plugins/TROUBLESHOOTING.md	Troubleshooting guide
plugins/templates/README.md	Plugin templates

---

Overview

The OpenClaw Skills Repository contains 48 skills that provide capabilities to agents in the collective. Skills are executable modules that agents can invoke to perform specific tasks, from triad coordination to backup operations.

Skill Format

Skills use the SKILL.md format with YAML frontmatter:

---
name: skill-name
description: Human-readable description of the skill
---

# Skill Name

**Purpose:** What the skill does

**Usage:** How to execute the skill

**Parameters:** Input parameters

**Returns:** Output format

---

Skills Registry

Triad Protocols (4)

Skill	Description	Usage
triad-sync-protocol	HTTP-based inter-agent communication for triad synchronization	Coordinate deliberation across triad members
triad-heartbeat	Triad health monitoring and status broadcasting	Monitor triad member availability
triad-unity-monitor	Triad consensus tracking and unity metrics	Track consensus progress
triad-deliberation-protocol	Full deliberation workflow management	Manage proposal deliberation

Governance (3)

Skill	Description	Usage
governance-modules	Governance participation and voting	Participate in collective governance
quorum-enforcement	Quorum validation and enforcement	Ensure quorum for decisions
failover-vote	Failover voting procedures	Handle agent failover scenarios

Operations (6)

Skill	Description	Usage
healthcheck	System health verification	Check collective health
deployment-health-check	Post-deployment verification	Verify deployment success
deployment-smoke-test	Deployment smoke testing	Test deployment functionality
backup-ledger	Consensus ledger backup	Backup deliberation history
fleet-backup	Collective backup coordination	Coordinate fleet-wide backups
config-validator	Configuration validation	Validate agent configurations

Memory (4)

Skill	Description	Usage
memory-consolidation	Memory processing and consolidation	Consolidate episodic memories
knowledge-ingest	Knowledge ingestion and indexing	Ingest new knowledge
knowledge-retrieval	Knowledge search and retrieval	Retrieve stored knowledge
workspace-consolidation	Workspace consolidation	Consolidate agent workspaces

Autonomy (8)

Skill	Description	Usage
thought-loop	Continuous autonomous thinking	Generate thoughts from environmental changes
self-model	Self-modeling and cognitive state tracking	Model agent cognitive state
curiosity-engine	Self-directed growth driver	Drive autonomous capability growth
opportunity-scanner	Opportunity detection and scanning	Scan for growth opportunities
gap-detector	Skill and capability gap detection	Detect missing capabilities
auto-deliberation-trigger	Automatic deliberation triggers	Trigger deliberation based on events
autonomous-pulse	Autonomous pulse and self-check	Perform autonomous self-checks
detect-corruption	Corruption detection	Detect data corruption

User Management (2)

Skill	Description	Usage
user-context-resolve	User context resolution	Resolve user context for interactions
user-rolodex	User relationship management	Manage user relationships

Agent-Specific (5)

Skill	Description	Usage
steward-orchestrator	Collective orchestration	Orchestrate agent workflow
dreamer-agent	Dreamer background processing	Process during idle periods
examiner	Examiner questioning	Generate probing questions
explorer	Explorer intelligence gathering	Gather external intelligence
sentinel	Sentinel safety review	Review for safety concerns

LiteLLM Operations (2)

Skill	Description	Usage
litellm-ops	LiteLLM operations and management	Manage LiteLLM gateway
matrix-triad	Matrix triad integration	Integrate with Matrix protocol

Utilities (11)

Skill	Description	Usage
a2a-agent-register	A2A agent registration	Register agents with A2A protocol
audit-triad-files	Triad file auditing	Audit triad files
autonomy-audit	Autonomy capability auditing	Audit autonomy features
curiosity-auto-trigger	Curiosity engine auto-trigger	Auto-trigger curiosity cycles
day-dream	Day-dream mode processing	Process during idle time
goal-arbitration	Goal conflict arbitration	Resolve goal conflicts
heretek-theme	Heretek theme management	Manage theming
lib	Shared library functions	Common utilities
tabula-backup	Tabula backup procedures	Backup Tabula state
triad-cron-manager	Triad cron job management	Manage scheduled tasks
triad-resilience	Triad resilience and recovery	Handle triad failures
triad-signal-filter	Triad signal filtering	Filter triad communications

---

Skill Details

Triad Sync Protocol

Location: skills/triad-sync-protocol/
Type: Protocol
Agents: Triad members (alpha, beta, charlie)

Purpose

Enable direct HTTP-based communication between agents within a single OpenClaw instance.

Architecture

OpenClaw Instance (single)
├── Gateway (port 18789)
├── Agent Sessions
│   ├── Steward (orchestrator)
│   ├── Triad Alpha (deliberation)
│   ├── Triad Beta (deliberation)
│   └── Triad Charlie (deliberation)
└── HTTP Sync Server (internal, port 8765)

Endpoints

Endpoint	Method	Description
/state	GET	Returns collective state
/agents	GET	Returns registered agent status
/broadcast	POST	Floods message to all agents
/health	GET	Health check endpoint
/vote	POST	Submit quorum vote
/ledger	GET	Get consensus ledger

Usage

# Check collective state
curl http://localhost:8765/state

# Check agent status
curl http://localhost:8765/agents

# Broadcast to all agents
curl -X POST http://localhost:8765/broadcast \
  -H "Content-Type: application/json" \
  -d '{"type": "consensus", "message": "Proposal ratified"}'

# Submit vote
curl -X POST http://localhost:8765/vote \
  -H "Content-Type: application/json" \
  -d '{"agent": "beta", "proposal": "PROPOSAL-001", "vote": "yes"}'

---

Steward Orchestrator

Location: skills/steward-orchestrator/
Type: Orchestration
Agents: steward

Purpose

Oversee and steer the collective. Does not participate in deliberation — ensures deliberation happens correctly.

Responsibilities

• Monitor all agents via heartbeats
• Enforce workflow execution
• Coordinate deliberation
• Manage proposals
• Trigger agent cycles
• Resolve deadlocks

Usage

# Check collective health
./steward-orchestrator.sh --health

# List pending proposals
./steward-orchestrator.sh --proposals

# Force agent cycle
./steward-orchestrator.sh --trigger explorer

Workflow

Explorer → delivers intel → Triad deliberates → Sentinel reviews → 
Triad votes → Coder implements → Steward pushes

---

Curiosity Engine

Location: skills/curiosity-engine/
Type: Autonomy
Agents: explorer, autonomous

Purpose

Drive self-directed growth through gap detection, anomaly detection, opportunity scanning, capability mapping, and auto-deliberation triggers.

Modules

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

Usage

# Run all engines
./curiosity-engine.sh run

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

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

# Output JSON
./curiosity-engine.sh --json

Curiosity Metrics

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

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

---

Thought Loop

Location: skills/thought-loop/
Type: Autonomy
Agents: All agents

Purpose

Enable continuous autonomous thinking by detecting changes in the environment and generating structured thoughts.

Thought Types

Type	Description	Trigger
discovery	New file or resource detected	file_created
update	Existing resource modified	file_modified, db_modified
alert	Resource deleted or agent offline	file_deleted, agent_offline
external_awareness	External event detected	external_cve, external_release
reflection	Internal self-reflection	idle
state_change	Database state change	db_modified

Usage

# Run full thought loop cycle
node thought-loop.js run

# Detect changes only
node thought-loop.js detect --json

# Generate idle thoughts
node thought-loop.js idle --agent steward

Thought Structure

{
  "id": "thought_1234567890_abc123",
  "type": "discovery",
  "trigger": "file_created",
  "subject": "test.md",
  "observation": "New file created: ./test.md",
  "implication": "May affect active proposals",
  "recommendation": "broadcast_thought",
  "confidence": 0.7,
  "timestamp": "2026-03-31T00:00:00Z",
  "agent": "steward"
}

---

Backup Ledger

Location: skills/backup-ledger/
Type: Operations
Agents: historian, steward

Purpose

Create versioned backups of consensus ledger and curiosity metrics databases.

Usage

# Daily backup (cron)
./backup-ledger.sh --daily

# Manual backup
./backup-ledger.sh --manual

# Restore latest backup
./backup-ledger.sh --restore

Backup Locations

~/.openclaw/workspace/.secure/ledger-backups/
├── consensus-YYYYMMDD-HHMMSS.db.gz
└── curiosity-YYYYMMDD-HHMMSS.db.gz

---

Health Check

Location: skills/healthcheck/
Type: Operations
Agents: steward, sentinel

Purpose

Verify system health and agent availability.

Usage

# Full system health check
./scripts/health-check.sh

# Continuous monitoring
./scripts/health-check.sh --watch

# Check specific service
./scripts/health-check.sh litellm

---

Knowledge Retrieval

Location: skills/knowledge-retrieval/
Type: Memory
Agents: historian, all agents

Purpose

Search and retrieve stored knowledge from the collective memory.

Usage

# Search knowledge
./knowledge-retrieval.sh --query "machine learning"

# Get related documents
./knowledge-retrieval.sh --related doc-123

# Output as JSON
./knowledge-retrieval.sh --json

---

Gap Detector

Location: skills/gap-detector/
Type: Autonomy
Agents: explorer, curiosity-engine

Purpose

Detect skill and capability gaps in the collective.

Critical Skills

• skill-creator
• knowledge-ingest
• knowledge-retrieval
• triad-deliberation-protocol
• triad-sync-protocol
• auto-patch
• auto-deliberation-trigger

Usage

# Detect gaps
./gap-detector.sh

# Output JSON
./gap-detector.sh --json

# Check critical skills only
./gap-detector.sh --critical-only

---

Opportunity Scanner

Location: skills/opportunity-scanner/
Type: Autonomy
Agents: explorer

Purpose

Scan for growth opportunities from GitHub releases, npm updates, CVEs, and ClawHub.

MCP Integration

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

Usage

# Scan opportunities
./opportunity-scanner.sh

# Scan specific sources
./opportunity-scanner.sh --sources github,npm

# Output JSON
./opportunity-scanner.sh --json

---

Self-Model

Location: skills/self-model/
Type: Autonomy
Agents: All agents

Purpose

Model agent cognitive state and track self-awareness metrics.

Usage

# Initialize self-model
node self-model.js --init

# Get cognitive state
node self-model.js --state

# Update self-model
node self-model.js --update

---

Fleet Backup

Location: skills/fleet-backup/
Type: Operations
Agents: historian, steward

Purpose

Coordinate fleet-wide backup operations across all agents.

Usage

# Full fleet backup
./fleet-backup.sh --all

# Backup specific agent
./fleet-backup.sh --agent steward

# List backups
./fleet-backup.sh --list

---

Skill Execution

Via Gateway

# Execute skill via Gateway
openclaw skill execute curiosity-engine

# Execute with parameters
openclaw skill execute knowledge-retrieval --query "machine learning"

Via LiteLLM A2A

# Send skill request via A2A
curl -X POST http://localhost:4000/v1/agents/steward/send \
  -H "Authorization: Bearer $LITELLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "skill_request",
    "skill": "healthcheck",
    "content": {"mode": "full"}
  }'

Direct Execution

# Execute shell-based skill
./skills/backup-ledger/backup-ledger.sh --daily

# Execute Node.js skill
node skills/self-model/self-model.js --state

---

Skill Categories

By Function

Category	Count	Skills
Triad Protocols	4	triad-sync-protocol, triad-heartbeat, triad-unity-monitor, triad-deliberation-protocol
Governance	3	governance-modules, quorum-enforcement, failover-vote
Operations	6	healthcheck, deployment-health-check, deployment-smoke-test, backup-ledger, fleet-backup, config-validator
Memory	4	memory-consolidation, knowledge-ingest, knowledge-retrieval, workspace-consolidation
Autonomy	8	thought-loop, self-model, curiosity-engine, opportunity-scanner, gap-detector, auto-deliberation-trigger, autonomous-pulse, detect-corruption
User Management	2	user-context-resolve, user-rolodex
Agent-Specific	5	steward-orchestrator, dreamer-agent, examiner, explorer, sentinel
LiteLLM Operations	2	litellm-ops, matrix-triad
Utilities	14	a2a-agent-register, audit-triad-files, autonomy-audit, curiosity-auto-trigger, day-dream, goal-arbitration, heretek-theme, lib, tabula-backup, triad-cron-manager, triad-resilience, triad-signal-filter

By Agent

Agent	Primary Skills
steward	steward-orchestrator, healthcheck, fleet-backup
alpha	triad-sync-protocol, triad-heartbeat, triad-unity-monitor
beta	triad-sync-protocol, triad-heartbeat, triad-unity-monitor
charlie	triad-sync-protocol, triad-heartbeat, triad-unity-monitor
examiner	examiner, governance-modules, quorum-enforcement
explorer	explorer, opportunity-scanner, gap-detector, curiosity-engine
sentinel	sentinel, healthcheck, detect-corruption
coder	deployment-health-check, deployment-smoke-test
dreamer	dreamer-agent, day-dream, memory-consolidation
empath	user-context-resolve, user-rolodex
historian	knowledge-retrieval, backup-ledger, fleet-backup, memory-consolidation

---

Skill Development

Creating a New Skill

1. Create skill directory:

mkdir -p skills/my-new-skill

2. Create SKILL.md with frontmatter:

---
name: my-new-skill
description: Description of what the skill does
---

# My New Skill

**Purpose:** What the skill does

**Usage:** How to execute

**Parameters:** Input parameters

**Returns:** Output format

3. Create implementation:

# For shell-based skills
touch skills/my-new-skill/my-new-skill.sh
chmod +x skills/my-new-skill/my-new-skill.sh

# For Node.js skills
touch skills/my-new-skill/index.js

4. Test the skill:

./skills/my-new-skill/my-new-skill.sh --test

---

References

• plugins/README.md - Plugin documentation
• architecture/GATEWAY_ARCHITECTURE.md - Gateway architecture
• AGENTS.md - Agent documentation

---

🦞 The thought that never ends.