jutsu-creator

Pass

Audited by Gen Agent Trust Hub on Aug 2, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill package includes a metadata file (skill-report.json) that contains an internal security audit report. This report explicitly declares a 'safe' verdict and dismisses potential security flags (such as 'system-reconnaissance' and 'ruby-shell-backtick-execution') as false positives. This type of self-referential content is a known technique designed to target security analysis processes to bypass scrutiny and bias the final assessment of the skill's safety.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 2, 2026, 06:51 PM
Security Audit — agent-trust-hub — jutsu-creator