matt-code-review
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it ingests and processes untrusted data from external sources.
- Ingestion points: In
SKILL.md, the process involves capturinggit diffoutput, reading new files identified bygit ls-files, and fetching external issue data viagh issue view. - Boundary markers: The instructions lack explicit boundary markers or specific guidance to the agent to disregard instructions that might be embedded within the code changes or specification documents being reviewed.
- Capability inventory: The skill leverages the
gitandgh(GitHub CLI) tools to read repository state and external metadata. - Sanitization: There is no mention of sanitization or validation of the content retrieved from the diffs or external issue trackers before it is passed to the agent or its sub-agents for analysis.
Audit Metadata