eval-harness

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill describes a legitimate software development methodology focused on testing and evaluation.
  • [COMMAND_EXECUTION]: The skill uses bash commands such as 'npm test' and 'grep' for deterministic code grading. These commands are typical for the described purpose of running software tests and do not involve suspicious remote downloads or privilege escalation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 07:55 AM
Security Audit — agent-trust-hub — eval-harness