skills/openai/plugins/evaluate-plugin/Gen Agent Trust Hub

evaluate-plugin

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • Command Line Interface Integration: The skill uses the plugin-eval utility to perform analysis, comparisons, and reporting. It passes local directory paths and user-specified strings as arguments to this tool to generate performance metrics and audits.
  • Local Directory Analysis: The skill is designed to read and audit content from a specified plugin root directory, specifically looking for .codex-plugin/plugin.json. This access is necessary for providing detailed feedback on manifest findings and skill strengths.
  • Indirect Prompt Injection Surface: The skill ingests data from external plugin files (ingestion point: <plugin-root>) and processes it using the plugin-eval tool. While boundary markers like quotes are used for user requests, the skill processes instructions embedded within the analyzed plugin files to provide a comprehensive audit.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 09:39 AM
Security Audit — agent-trust-hub — evaluate-plugin