release
Pass
Audited by Gen Agent Trust Hub on May 28, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill orchestrates standard development tools such as git, npm, pytest, and go test to perform release tasks. It also executes local repository scripts like scripts/ci-local-release.sh and scripts/verify-release-ci.sh, which are expected parts of a project's automation infrastructure.
- [PROMPT_INJECTION]: The skill ingests untrusted data from git commit messages to generate changelogs (Indirect Prompt Injection surface). This is mitigated by a mandatory 'Step 8: User review' phase where the agent must present all proposed changes to the user for explicit confirmation before writing to disk or creating commits.
- [DATA_EXFILTRATION]: The skill explicitly restricts its operations to the local environment, stating it does not push code, publish packages, or trigger remote CI actions directly. This design prevents unauthorized data exfiltration.
Audit Metadata