ui-code-auditor
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code files (such as .tsx, .jsx, and .erb) which could contain malicious comments or text patterns intended to subvert the agent's audit logic or safety guidelines.
- Ingestion points: Project source files are ingested during Step 4 of the Review Process in
SKILL.mdusing theReadandGreptools. - Boundary markers: The skill does not implement explicit delimiters or system instructions to disregard instructions embedded within the source files being audited.
- Capability inventory: The skill uses
Bash(forfind),Grep,Read, andGlobtools to perform its auditing tasks. - Sanitization: No sanitization or pre-processing of file content is performed to filter out potential injection attacks before the agent evaluates the code against its rules.
Audit Metadata