changelog-generator
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze external data from git commit histories, which could contain malicious instructions if the repository contains commits from untrusted sources.
- Ingestion points: Git commit history and messages are read from the project repository (referenced in SKILL.md).
- Boundary markers: The instructions do not specify any boundary markers or delimiters to help the agent distinguish between the commit data and its own processing instructions.
- Capability inventory: The skill encourages the agent to read repository history and write output to a local file (
CHANGELOG.md). - Sanitization: There are no instructions or mechanisms defined to sanitize or filter the commit messages before they are processed by the LLM.
Audit Metadata