release
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes multiple git commands (branch validation, log retrieval, commit, tag, and push) to manage the repository release cycle.
- [EXTERNAL_DOWNLOADS]: Interacts with the GitHub API using the gh CLI to fetch pull request and author metadata for changelog generation.
- [PROMPT_INJECTION]: The skill ingests untrusted data from commit messages in the git log. 1. Ingestion points: git log output. 2. Boundary markers: None. 3. Capability inventory: git push and file-write operations. 4. Sanitization: None. This indirect prompt injection surface is mitigated by the mandatory user review step before any changes are finalized.
- [SAFE]: Implements a mandatory human-in-the-loop workflow using AskUserQuestion, ensuring explicit user oversight for all irreversible operations.
Audit Metadata