check-harness
Pass
Audited by Gen Agent Trust Hub on Jun 29, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from the local environment.
- Ingestion points: The skill reads project metadata (e.g.,
README.md,package.json,CLAUDE.md) and user message text from session logs (referenced inreferences/questionnaire.md). - Boundary markers: The instructions lack explicit boundary markers or instructions to treat the read content strictly as data, which may allow malicious instructions embedded in project files or past logs to influence agent behavior.
- Capability inventory: The skill utilizes
Bash,Write, andReadtools, providing a significant capability tier that could be exploited via injection. - Sanitization: There is no evidence of sanitization or validation of the content retrieved from project files or logs before processing.
Audit Metadata