Stats
Actions
Tags
From codex-review
Get a consensus code review by running a deliberation between Claude and Codex (a separate AI model). Both models independently review the code, then debate their findings until they agree. Use this skill whenever the user asks for a code review, wants feedback on their changes, says "review my code", "review this", "what do you think of these changes", "check my work", "does this look right", or wants a second opinion on code. Also trigger when the user mentions "codex review", asks to "run a review", or wants another model to look at their code. Even casual requests like "any issues with this?" or "how does this look?" should trigger this skill when there are code changes in the working tree.
How this skill is triggered — by the user, by Claude, or both
Slash command
/codex-review:codex-reviewThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Get a high-confidence code review through deliberation between two independent models. Claude and Codex each review the code separately, then go back and forth to challenge, validate, and reconcile their findings. The result is a consensus review where both models agree — catching more issues and filtering out false positives.
Get a high-confidence code review through deliberation between two independent models. Claude and Codex each review the code separately, then go back and forth to challenge, validate, and reconcile their findings. The result is a consensus review where both models agree — catching more issues and filtering out false positives.
This plugin bundles the Codex MCP server. You need codex installed and available on your PATH.
Use the branch the user specifies. If none is specified, use the repository's default branch (typically main or master).
Use the Agent tool to spawn a subagent with the instructions from ${CLAUDE_PLUGIN_ROOT}/agents/reviewer.md. Pass along:
The subagent handles the full deliberation loop with Codex and returns a consensus review.
The agent returns a consensus report with findings both models agreed on. Present it to the user as-is.
Save the THREAD_ID from the response — you can use it for follow-ups with Codex via mcp__codex__codex-reply.
If the user wants to discuss a finding, ask about alternatives, or push back on a recommendation, call mcp__codex__codex-reply with the saved threadId and the user's follow-up question.
npx claudepluginhub scristobal/claude-plugins --plugin codex-reviewRuns cross-model code reviews using the external Codex CLI tool from a Claude session. Catches bugs that single-model self-review would miss by leveraging a different reviewer architecture.
Cross-model review using OpenAI Codex to independently verify plans or code diffs, iterating up to 5 rounds. Useful for architecture decisions, non-trivial refactors, and critical config changes.
Runs code reviews via external LLMs (OpenAI Codex CLI, Google Gemini CLI) on uncommitted changes, branch diffs, or commits using git and bash.