rea-changelog-update
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands within the local environment. Evidence includes the use of
node scripts/release-notes.mjsfor data inventory, recording, and rendering, as well asgh pr editto update GitHub Pull Request descriptions (file: SKILL.md).\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Git history and GitHub metadata, presenting a surface for potential injection attacks.\n - Ingestion points: Git commit history and author metadata, GitHub Pull Request titles/descriptions, and issue tracker details (file: SKILL.md).\n
- Boundary markers: The instructions do not define delimiters to wrap or isolate the processed metadata during changelog drafting (file: SKILL.md).\n
- Capability inventory: The skill can write to the local filesystem and modify remote Pull Request bodies via the GitHub CLI (file: SKILL.md).\n
- Sanitization: While the skill advises manual review of diffs and verification of contributor handles, it does not specify technical sanitization or escaping of the ingested text (file: SKILL.md).
Audit Metadata