design-system-enforcer
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingest untrusted source code from the local repository to audit styling and component usage, which creates a surface where malicious instructions in comments or code strings could attempt to influence the agent behavior.
- Ingestion points: Source code files (TS, TSX, JS, JSX, CSS, SCSS) are read into context via the
scripts/scan_style_drift.pyscript and manual inspection. - Boundary markers: No delimiters or explicit instructions are used to isolate the analyzed content from the agent's control logic.
- Capability inventory: The skill is limited to file reading and pattern matching; no network exfiltration or file-writing capabilities are present in the provided files.
- Sanitization: The script reads raw text content without sanitizing or escaping potential injection patterns.
Audit Metadata