test-evidence-review
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from repository files which could contain malicious instructions.
- Ingestion points: Phase 1 in
references/claude-skill.mdreads stories and acceptance criteria from the repository verbatim. Phase 2 reads test source files fromAssets/Tests/EditMode/**andAssets/Tests/PlayMode/**, as well as NUnit XML results. - Boundary markers: No boundary markers, delimiters, or instructions to disregard embedded commands are defined in the workflow phases.
- Capability inventory: The skill utilizes
Read,Glob,Grep, andWritetools. It has the capability to write review reports to theproduction/qa/directory when the--recordflag is used. - Sanitization: There is no evidence of sanitization, escaping, or validation of the content retrieved from external files before it is processed by the agent's logic.
Audit Metadata