Add clawheart-security skill#41
Conversation
Drives ClawHeart's local AI security runtime (https://clawheart.live) via the `clawheart` CLI to: - Run 80-check AI security scans - Audit installed skills for malicious behavior - Inventory installed AI agents (Claude Code / Codex / Cursor / OpenClaw / Continue / Gemini CLI / Windsurf) - List MCP servers per agent - Inspect third-party LLM provider profiles All commands emit `{ok, data, error}` JSON; the skill specifies how to surface results in natural language. Validator: local `node scripts/validate-skills.mjs` ✓
✅ Validate Skills — PASSEDSummary: FAIL=0 · WARN=0 Output (last 200 lines) |
🛡️ Skills Security Scan — FAILSummary: FAIL=2 · WARN=7 · FILES=1 ❌ FAIL blocks merge. Fix the items below. ❌ Failuresclawheart-security
Guidance: Blocks download-and-execute. Vendor scripts or verify pinned downloads + checksums.
Guidance: Blocks download-and-execute. Vendor scripts or verify pinned downloads + checksums.
|
Submit a Skill
Thank you for contributing a skill to the Goose community!
📛 Skill Info
📁 Structure Checklist
🔐 Security Acknowledgement
📝 Additional Notes
This skill is a pure-instruction SKILL.md — it tells the agent which `clawheart` subcommand to call for each user intent (scan / skills audit / agents list / MCP / providers / status / onboarding), and how to render the structured JSON output. No scripts, no templates, no binaries.
About ClawHeart: ClawHeart Desktop is a local-first AI security runtime that proxies AI agent traffic, audits installed skills, and inventories agent + MCP + provider configuration on the user's machine. The CLI (`clawheart-cli`, ≤2MB) is the JSON-first entry point this skill drives.
Source: https://github.com/tjsdyy/clawheartv2 · CLI release: https://github.com/tjsdyy/clawheartv2/releases/latest
Validator locally: `node scripts/validate-skills.mjs` ✓