improve-skill
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill is a maintenance tool for improving other agent skills, and its operations are aligned with its documented purpose.
- [EXTERNAL_DOWNLOADS]: Fetches documentation from well-known official sources (package registries, API documentation) to verify accuracy of technical claims within the analyzed skill. These operations target trusted sources and do not escalate the verdict.
- [COMMAND_EXECUTION]: Modifies local files using Edit and Write tools to apply user-approved improvements. These actions require explicit user confirmation before being executed.
- [PROMPT_INJECTION]: The skill processes untrusted input in the form of existing skill files. Ingestion points: Reads skill files from user-provided paths in Phase 0 and fetches external content in Phase 2b. Boundary markers: None identified; content is analyzed as data. Capability inventory: Uses Edit and Write for file modifications and Task for specialized subagent linting. Sanitization: No explicit sanitization of input skill content is performed before analysis.
Audit Metadata