release
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes shell commands for git repository management, environment validation, and versioning analysis.
- [COMMAND_EXECUTION]: Runs a local deployment script (
./scripts/release.sh) to perform build and release operations. This is a primary function of the skill. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from git commit logs to generate human-readable changelog entries.
- Ingestion points: Git commit history and the project's
CHANGELOG.mdfile. - Boundary markers: Not explicitly present in instructions when processing external data.
- Capability inventory: Includes file system modifications, git commit/push operations, and shell script execution.
- Sanitization: The risk is mitigated by explicit instructions for the agent to rewrite commit messages for users and a mandatory human confirmation step before the changelog or release proceeds.
Audit Metadata