rw-research-design
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a detailed methodology for research design with clear boundaries. It explicitly instructs the agent not to access private local workspaces or personal directories.
- [SAFE]: External references point to well-known academic and clinical research standard organizations (e.g., equator-network.org, consort-spirit.org, strobe-statement.org).
- [SAFE]: Includes a validation script (
scripts/self_check.py) designed to ensure skill integrity. The script scans the skill's own files for forbidden strings like '/Users/' or 'private-workspace/' to prevent accidental hardcoding of sensitive local paths. - [SAFE]: The instructions emphasize source discipline, instructing the agent to distinguish between confirmed facts and inferences and to avoid generating non-existent data or citations.
Audit Metadata