asif999/openclaw-config
Summary
OpenClaw agent configuration, workspace bootstrap files, and skill library. Serves as backup and version control for the complete OpenClaw agent setup — agents, models, MCP bridges, port registry, and skill definitions.
Included Assets
- Agent registry (
AGENTS.md) — 17 active agents with assigned models and purposes - Agent configuration (
openclaw.json) — models, ACP settings, allowed agents, concurrency limits - Long-term memory (
MEMORY.md) — persistent facts, preferences, and decisions across sessions - Architecture docs (
docs/) — stack diagram, port registry, config change log, deferred skill TODOs - Skill library (
skills/) — 9 skill definitions with install instructions, commands, and usage patterns
Skills
| Skill | Description |
|---|---|
1password | 1Password CLI (op) — sign-in, secret read/inject, op run |
apple-notes | Apple Notes via memo CLI — create, view, edit, delete, search |
apple-reminders | Apple Reminders via remindctl — list, add, complete, delete |
bear-notes | Bear notes via grizzly CLI — create, search, tag management |
blogwatcher | RSS/Atom feed monitoring — scan, list articles, mark read |
blucli | BluOS CLI — Bluesound/NAD player control, grouping, TuneIn |
bluebubbles | iMessage via BlueBubbles — send, react, edit, unsend, reply |
camsnap | RTSP/ONVIF camera capture — snapshots, clips, motion watch |
canvas | HTML rendering on connected OpenClaw nodes via canvas tool |
How to Use
This repository is the source of truth for the OpenClaw workspace. To restore or replicate the setup:
- Copy
openclaw.jsonto~/.openclaw/openclaw.json - Review
AGENTS.mdfor the active agent roster and model assignments - Populate
MEMORY.mdwith project-specific facts and preferences - Apply port assignments from
docs/PORT-REGISTRY.mdas needed - Reference individual
skills/*.SKILL.mdfiles for CLI setup and usage per capability
For config changes: follow the protocol in docs/CHANGELOG-openclaw-json.md — use openclaw config set, record the change, and commit both files together.
Notes
- The stack uses 19 agents across 26 layers with 8 concurrent session slots
- Primary LLM: MiniMax M2.7 via OpenAI-compatible API
- Fallback: Claude Sonnet 4.6 via Anthropic
- MCP bridge layer connects to 8 external services (GitHub, Linear, Sourcegraph, Outline, etc.)
- All port assignments are centralized in
docs/PORT-REGISTRY.md— update there before changing any service port - Several skill layers have unresolved TODOs listed in
docs/TODO-SKILLS.md— these require actual first-use to complete audit
No comments yet.