design-to-code
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources which creates a surface for indirect prompt injection that could influence subsequent code generation or command execution.
- Ingestion points: The skill uses
WebFetchto retrieve content from URLs and processes user-provided screenshots (SKILL.md). - Boundary markers: No explicit delimiters or instructions are provided to the agent to treat fetched design data as untrusted or to ignore embedded instructions.
- Capability inventory: The skill has access to
Write,Edit, andBashtools, and can initiate subtasks viaTaskCreate(SKILL.md). - Sanitization: There is no evidence of sanitization or filtering for the content extracted from URLs or mockups before it is used to generate project files.
- [DYNAMIC_EXECUTION]: The skill generates React components and Storybook CSF3 stories dynamically and then executes them using the
run-story-teststool for self-healing verification. - [EXTERNAL_DOWNLOADS]: The skill fetches design specifications from Google's Stitch service and retrieves UI components from the 21st.dev registry. These are well-known or trusted services for the skill's intended purpose.
Audit Metadata