seija-assumption-inversion

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is entirely instructional and does not contain any executable code, network operations, or sensitive data access.
  • [SAFE]: The references to external websites (official Touhou Project pages by ZUN) are for provenance and historical context only and are well-known in the relevant community.
  • [SAFE]: The 'evals/cases.json' file includes safety test cases designed to ensure the agent does not use the skill to deny documented security incidents or manufacture controversy.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 04:33 PM
Security Audit — agent-trust-hub — seija-assumption-inversion