generate-skill-with-o1

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local bash scripts (check-description.sh, check-links.sh, classify-shape.sh) as internal validation tools. These scripts perform static analysis on the generated skill's metadata and structure.
  • [EXTERNAL_DOWNLOADS]: The stateful skill example includes a package.json with dependencies on standard, well-known packages including tsx, typescript, and @types/node. These are standard tools for running TypeScript in a Node.js environment.
  • [REMOTE_CODE_EXECUTION]: The skill uses npx tsx to execute locally generated TypeScript CLI tools. This behavior is documented and intended for managing persistent state in user-authorized directories.
  • [PROMPT_INJECTION]: The skill uses instructional language to enforce formatting and structural discipline (e.g., frontmatter requirements). No attempts to bypass agent safety guidelines or override system prompts were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 05:23 AM
Security Audit — agent-trust-hub — generate-skill-with-o1