taste-stitch-design-taste
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to ingest user-supplied "vibe descriptions" and incorporate them into a generated
DESIGN.mdfile. This generated file is intended to serve as a set of instructions for another AI agent (Google Stitch), which creates a potential surface for indirect prompt injection where malicious input could influence downstream agent behavior. - Ingestion points: User-provided project intent and atmosphere descriptions processed in the "Analysis & Synthesis Instructions" section of
SKILL.md. - Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to disregard potentially malicious commands embedded in user-supplied vibe descriptions.
- Capability inventory: The skill generates complex instructional markdown used to steer the visual and functional output of the Google Stitch agent.
- Sanitization: There are no documented steps for sanitizing or escaping the user-provided descriptions before they are interpolated into the design system rules.
Audit Metadata