agent-instructions-evaluator

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, persistence mechanisms, or unauthorized file access detected.
  • [EXTERNAL_DOWNLOADS]: The skill requires the 'pyyaml' package via 'requirements.txt'. This is a well-known, standard library from the official PyPI registry and is used safely for parsing configuration files.
  • [COMMAND_EXECUTION]: Utility scripts for static analysis perform structural inspection of Python, JSON, and YAML files. These scripts use the 'ast' module and 'yaml.safe_load' to ensure that content is analyzed without being executed at runtime.
  • [DATA_EXFILTRATION]: No network-enabled operations or exfiltration patterns were found. The tool logic is contained within the local environment.
  • [PROMPT_INJECTION]: The skill serves as a quality assessment tool for agent instructions and does not contain any instructions aimed at bypassing model safety guidelines or exfiltrating data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 12:47 PM
Security Audit — agent-trust-hub — agent-instructions-evaluator