design-system-reconnaissance
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes project configuration and source files (such as CSS variables, Tailwind configurations, and JSON tokens) to identify design patterns. While these external files represent an ingestion surface for untrusted data, the skill mitigates risk by restricting the agent's write capabilities solely to a documentation file (
.readyvibe/design.md) and explicitly forbidding modifications to product code. - [SAFE]: The instructions are strictly limited to reconnaissance and documentation tasks. There are no signs of command execution, network exfiltration, persistence mechanisms, or credential harvesting. The skill operates within a restricted scope focused on visual consistency and design token identification.
Audit Metadata