compliance-matrix
Pass
Audited by Gen Agent Trust Hub on Feb 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The requirement extraction logic is vulnerable to indirect prompt injection.
- Ingestion points: The
extractRequirementsfunction inSKILL.mddirectly interpolates untrusted RFP text (rfp.description) into the AI prompt template. - Boundary markers: There are no delimiters or specific instructions (e.g., 'ignore embedded commands') to separate the untrusted RFP content from the system instructions.
- Capability inventory: The skill has the capability to read files using
Read,Grep, andGlobtools and can perform database writes via Convex mutations (saveComplianceMatrix). - Sanitization: The skill does not perform any sanitization or validation of the input RFP text before sending it to the AI provider.
Audit Metadata