draft-release
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from git commit history to generate human-readable changelog entries and PR descriptions.
- Ingestion points: Git commit subjects and full bodies are retrieved via
git logcommands in Step 3 and Step 4 ofSKILL.md. - Boundary markers: Absent. There are no explicit instructions or delimiters used to ensure the agent ignores malicious instructions that might be embedded in commit messages.
- Capability inventory: The skill possesses
WriteandEditcapabilities for the local file system (CHANGELOG.md) and usesBash(gh *)to interact with the GitHub API for PR creation. - Sanitization: Absent. The skill relies on the agent's summarization logic to interpret commit messages, which lacks formal sanitization or filtering of potential injection strings within the history data.
Audit Metadata