stitch-ui-react-components
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external design data which may contain hidden instructions.\n
- Ingestion points: External design HTML and visual assets retrieved via signed URLs.\n
- Boundary markers: Lacks specific isolation markers to prevent the agent from following instructions embedded in design files.\n
- Capability inventory: Includes shell execution and file write capabilities.\n
- Sanitization: Masks personal information (PII) but does not perform sanitization for prompt injection.\n- [COMMAND_EXECUTION]: Uses the skill's own
scripts/fetch-stitch.shto download assets.\n- [EXTERNAL_DOWNLOADS]: Fetches design files from external URLs and references resources from the google-labs-code GitHub organization.
Audit Metadata