agent-rules
Rules and Knowledge to work better with agents such as Claude Code or Cursor
Rules and Knowledge to work better with agents such as Claude Code or Cursor Live-indexed from GitHub; open the repo to vet quality before installing anything.
Works with: OpenClaw · Claude Code · Cursor · Any MCP agent · v0.live · MIT · by steipete
SKILL.md preview
# agent-rules Community index (GitHub): https://github.com/steipete/agent-rules ## Triggers Looking for reusable code workflows. ## Workflow 1. Read the project README. 2. Extract the reusable pattern. 3. Adapt it into a local SKILL.md. ## Guardrails - Audit code before running. - Pin versions; review permissions.
Install in one command
1. Create the skill folder: mkdir -p skills/gh-1000814478
2. Install: curl -fsSL "https://skillforge-aniruddha-adaks-projects.vercel.app/api/skills/gh-1000814478/skill.md" -o skills/gh-1000814478/SKILL.md
3. Mention it in SOUL.md or BOOTSTRAP.md so OpenClaw loads it on boot.
curl -fsSL "https://skillforge-aniruddha-adaks-projects.vercel.app/api/skills/gh-1000814478/skill.md" -o skills/gh-1000814478/SKILL.md
Why this score
3/3 required SKILL.md sections, 320 chars
4 declared capabilities
Works with openclaw, claude-code, cursor, generic-mcp
MIT — reuse freely
5684 GitHub stars
Risk surface: network-egress, external-api
Audit seal
SHA-384(prev ‖ canonicalJson)
06ac2e68fca85f3311f02edb50f98f0a9a15e3d4c146277cc6f02e68c90adc65d574fa32c063c9456b3bb6369e89877a
chain head e800631e36483c866a9a2ba7ae87b272…
Replay: POST {ids} to /api/seals/verify