stitch-ui-designer

Pass

Audited by Gen Agent Trust Hub on Feb 22, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (LOW): The skill defines a workflow that ingests untrusted user data to drive automated tool actions, creating a vulnerability surface.
  • Ingestion points: User-provided design requests (e.g., "Design a dashboard for X") are directly passed as arguments to orchestrator tools like stitch-ui-design-spec-generator and stitch-ui-prompt-architect in SKILL.md and references/workflows.md.
  • Boundary markers: There are no implementation details or instructions to use delimiters (like XML tags or block quotes) to separate user input from the agent's internal prompt logic.
  • Capability inventory: The skill is granted powerful capabilities including stitch*:* (project management and UI generation), Write (file system modification), and web_fetch (external network operations).
  • Sanitization: The instructions lack any sanitization, filtering, or validation steps for the user-supplied strings before they are incorporated into the design generation prompt.
  • Safety Violation: The workflow explicitly mandates that the agent "ALWAYS execute immediately (no confirmation loop)", which removes the critical safety check of human verification for high-impact tool executions.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 22, 2026, 02:23 PM