diff
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs analytical tasks by reading and comparing design tokens from local files. It does not perform network operations, exfiltrate data, or execute untrusted code.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface by ingesting untrusted data from external sources.
- Ingestion points: User-provided file paths via $ARGUMENTS and brand identifiers from .design-lock.
- Boundary markers: Absent; there are no instructions to the agent to ignore instructions embedded in the target design files.
- Capability inventory: Local file system read access across multiple paths.
- Sanitization: Absent; content is extracted for properties without pre-filtering. As the skill's purpose is purely analytical and it lacks network or write capabilities, the risk is minimal.
Audit Metadata