release-notes
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to interact with the local repository and GitHub API.
- Evidence: Executes
git log,git tag,git push, and variousgh(GitHub CLI) commands to retrieve commit history and manage releases. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub which could contain instructions designed to influence the agent's behavior during the drafting phase.
- Ingestion points: Fetches PR and issue metadata (titles, bodies) using
gh pr listandgh issue listinSKILL.md. - Boundary markers: No explicit boundary markers or instructions to ignore embedded prompts are defined when processing these external strings.
- Capability inventory: Has the ability to create GitHub releases (
gh release create) and push git tags (git push) which could be abused if the agent is manipulated by injected content. - Sanitization: No specific sanitization or filtering of external content is mentioned beyond the thematic classification of items.
Audit Metadata