OpenAI Codex security directory

Plugin Eval

Published by OpenAI Codex

High risk

Evaluates local Codex plugins and skills through guided review and benchmarking. Explain scores, identify improvements and measure actual token usage. Turn findings into rewrite briefs or add custom evaluation criteria.

Security assessment

Plutonium assessed this plugin for permissions, capabilities, and security-relevant behavior.

  • Evaluation and benchmarking create local reports, configurations and worktrees: Can read, create, or change files on the local filesystem or in a selected project.
  • Evaluation reads local plugin, skill and repository content: Can read private account, workspace, repository, message, analytics, or file data.
  • Benchmark mode launches real Codex executions against the evaluated target: Can execute code, commands, scripts, unrestricted queries, or dynamic gateway actions.

Available capabilities

This add-on exposes 5 tools or capabilities.

  • evaluate-plugin
  • evaluate-skill
  • improve-skill
  • metric-pack-designer
  • plugin-eval

The interactive security report will load automatically.