rw-statistics-audit
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references official scientific reporting standards from well-known organizations such as EQUATOR Network, STROBE, and CONSORT/SPIRIT. These are trusted resources within the scientific community.
- [COMMAND_EXECUTION]: The skill includes a local validation script (scripts/self_check.py) intended for self-testing the skill's integrity and ensuring no hardcoded local paths are present. This script is static and does not process external or untrusted input.
- [PROMPT_INJECTION]: The skill processes user-provided research materials. The instructions include explicit boundaries to distinguish user input from public facts and prevent the generation of fabricated data. Evidence: (1) Ingestion points: User-provided text, files, and links (SKILL.md). (2) Boundary markers: Present in the 'Source Discipline' section (SKILL.md). (3) Capability inventory: Network access for verification and local execution of a static self-check script. (4) Sanitization: Not explicitly defined for external content, but operations are restricted to statistical auditing without code generation from user input.
Audit Metadata