finalize
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references external utilities including Google's
release-pleaseand Microsoft'sWazafor release automation and CI evaluation. - [COMMAND_EXECUTION]: Executes standard development lifecycle commands such as build, lint, typecheck, and test as part of the release verification process.
- [INDIRECT_PROMPT_INJECTION]: The skill consumes untrusted data from the git history and project files to generate release notes and reports.
- Ingestion points: Reads git log and conventional commits as part of the release workflow in
SKILL.md. - Boundary markers: No explicit delimiters are used to separate untrusted commit messages from the system instructions.
- Capability inventory: Shell command execution for build/test steps and interaction with release tools.
- Sanitization: No specific sanitization or escaping of commit message content is described prior to processing.
Audit Metadata