Scan Skill Endpoint
Grade an OpenClaw / Agent Skill in a GitHub repo — the capability surface a repo scan misses: the auto-exec allowed-tools grant, always-loaded- description injection, lifecycle-hook escalation, and env-exfil in bundled scripts. coverage.surface = openclaw. E.g. /public/scan/skill/owner/repo. The behavioral sandbox (plan skill) clones the repo and runs its hooks and scripts; the signed observation comes back in behavioral.
Request
- Base URL: https://agentavow.com (the document declares no server; this is the origin it was published from)
- URL: https://agentavow.com/api/v1/public/scan/skill/{owner}/{repo}
- Auth: none declared
Path parameters
Query parameters
Bypass cache and force a fresh scan
Force a fresh behavioral sandbox run: the skill is cloned and its lifecycle hooks and bundled scripts are run with canary credentials (implies a fresh scan). The run is also started automatically on the first scan.
Response
Successful Response
Changes
No recorded changes to this operation across all 1 revision of this API.