stitch-ui-prompt-architect
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill consists of markdown-based instructions for prompt engineering and UI design workflows. It provides clear guidance on avoiding the inclusion of sensitive information (PII, credentials) in prompts and strictly defines its scope as a text-generation assistant without direct execution capabilities. All external links point to official documentation for well-known UI frameworks or reputable developer platforms.- [INDIRECT_PROMPT_INJECTION]: While the skill ingests untrusted user requirements to generate UI prompts, it incorporates several security controls. It mandates data sanitization (replacing PII with demonstration values), uses structured sections ([Context], [Layout], [Components]) as natural boundaries, and limits its own output to text or file writing. The skill emphasizes verification and evidence-based reporting to minimize the risk of executing unintended instructions from the input data. Evidence: Ingestion points include user-provided UI requests; Boundary markers are provided via section headers; Capability inventory includes text generation and file-writing to next-prompt.md; Sanitization instructions are explicitly provided in SKILL.md and references/faq-deep.md.
Audit Metadata