prepare-release
Pass
Audited by Gen Agent Trust Hub on Jul 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes standard development utilities such as
git,npm install, andcargo generate-lockfileto automate repository management and dependency updates during a release. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and processes git commit history and existing changelog content which may contain untrusted data.
- Ingestion points: Content from
git logand theCHANGELOG.mdfile is read into the agent's context during Step 4. - Boundary markers: No explicit delimiters are specified for isolating the data ingested from git history.
- Capability inventory: The skill is capable of executing shell commands, writing files to the local repository, and pushing branches to a remote server.
- Sanitization: The skill relies on manual user review and confirmation of the drafted changelog as its primary security control; no automated filtering or sanitization of input data is defined.
Audit Metadata