design
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill architecture involves processing potentially untrusted external data, which creates a surface for indirect prompt injection attacks where malicious instructions could be embedded in the source material.
- Ingestion points: The workflow involves analyzing external codebases (DESIGN-SYSTEM mode), research transcripts, support tickets, and app reviews (RESEARCH mode), as well as Figma URLs (HANDOFF mode).
- Boundary markers: No specific delimiters or instructions to ignore embedded commands are included in the prompt templates for handling external data.
- Capability inventory: Analysis of all scripts and instructions shows no presence of dangerous capabilities such as file writing, network exfiltration tools, or arbitrary command execution.
- Sanitization: The instructions do not define sanitization or validation procedures for the ingested external content.
Audit Metadata