self-healing-ci
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill utilizes the gh-aw (GitHub Agentic Workflows) extension from the github organization (github/gh-aw) and templates referencing githubnext. These are recognized and authoritative sources for GitHub-integrated automation tools.\n- [COMMAND_EXECUTION]: Provided workflow examples use standard git and gh CLI commands to apply patches and re-run tests. These operations are scoped to the intended functionality of CI remediation and include verification gates to ensure code integrity.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface. 1. Ingestion points: CI logs and PR diffs are ingested for diagnosis (SKILL.md). 2. Boundary markers: No specific delimiters for log content are defined. 3. Capability inventory: The skill can propose PR comments, apply patches via git, and push commits (references/workflow-example.md). 4. Sanitization: The risk is mitigated by a mandatory re-run verification pass; a fix is only considered verified if the failing check passes after application, providing a behavioral gate against malicious instructions.
Audit Metadata