Launch
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill functions primarily as a planning coordinator, emphasizing safety through rollback designs, Go/No-Go gates, and explicit versioning rules. It is instructed to avoid direct deployment, deferring execution to other specialized agents.
- [COMMAND_EXECUTION]: The skill provides templates and guidance for standard repository operations using well-known tools such as git and the GitHub CLI (gh).
- [PROMPT_INJECTION]: Indirect prompt injection surface detected through the ingestion of external repository data. 1. Ingestion points: PR and commit history context retrieved from the Harvest agent during the 'Announce' phase. 2. Boundary markers: No explicit delimiters or safety instructions are defined to isolate untrusted commit messages. 3. Capability inventory: The skill generates executable commands for release management and documentation. 4. Sanitization: No explicit logic is documented to sanitize or validate the content of processed commit logs.
Audit Metadata