style

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill instructions and nested persona definitions in the style markdown files (e.g., styles/academia.md) are purpose-built for design system implementation and do not attempt to override the primary agent's safety constraints or reveal system prompts.
  • [DATA_EXFILTRATION]: No sensitive file access or network exfiltration patterns were identified. The skill writes generated tokens and documentation to local directories within the intended project structure.
  • [EXTERNAL_DOWNLOADS]: Documentation in SOURCES.md references external design resources on GitHub and designprompts.dev. These are included for transparency regarding data sources and are recognized as well-known/trusted services; the skill does not automate untrusted downloads at runtime.
  • [COMMAND_EXECUTION]: Although the skill includes Bash in its allowed tools, the procedural instructions do not contain arbitrary or high-risk shell commands. File operations are confined to the generation of design system assets.
  • [SAFE]: The skill is a self-contained utility for design token generation with no detected security issues.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 07:13 PM
Security Audit — agent-trust-hub — style