design-dna
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to fetch and process content from external, untrusted sources, specifically public URLs and screenshots provided by users. This architecture presents a surface for indirect prompt injection, where instructions hidden within the reference content could attempt to influence the agent's subsequent actions.
- Ingestion points: The skill accepts public URLs and user-supplied screenshots as primary inputs for design translation (documented in
SKILL.mdunder 'Input Priority' and 'Inspect Available Evidence'). - Boundary markers: There are no explicit boundary markers or 'ignore' instructions defined to prevent the agent from obeying commands found within the retrieved external data.
- Capability inventory: The agent possesses capabilities to browse external websites, generate executable HTML/CSS code, and capture/inspect screenshots of its own output.
- Sanitization: The skill does not specify any sanitization, filtering, or validation steps for the content retrieved from external URLs before it is processed by the model.
Audit Metadata