frontend-build
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is instructed to ingest and act upon content from design contracts (e.g.,
docs/design/<feature>/DESIGN.md) which could contain malicious instructions meant to influence the agent's behavior during the build process. - Ingestion points: The agent reads
docs/design/<feature>/DESIGN.mdand associated implementation issues/specs as the primary source of truth for visual and functional requirements. - Boundary markers: There are no explicit instructions for the agent to treat the design document content as untrusted or to isolate its contents from the execution environment.
- Capability inventory: The skill is authorized to run tests, typechecking, linting, and other repository tools, which could be misused if an injection in the design document successfully redirects the agent's actions.
- Sanitization: No sanitization or validation logic is specified for the external content before it is processed by the agent.
Audit Metadata