Run the full code review pipeline on your changes. Creates an agent team of parallel reviewers, aggregates findings, and fixes critical issues.
Use before implementing non-trivial features (researches approaches with Context7, Serper, GitHub MCPs)
Deep research using agent teams, web search, and scraping. Decomposes the topic into the questions that matter, researches in parallel via teammates, evaluates confidence, iterates until sufficient, and synthesizes with source attribution.
Review completed work against your plan and coding standards.
Reviews completed work against the original plan and coding standards. Invoked standalone via /review-code after completing a major project step. Not part of the code-review-pipeline dispatch — this is a plan-alignment reviewer.
Reviews whether documentation is up-to-date with code changes. Checks READMEs, architecture docs, API docs, changelogs, configuration docs, and inline docstrings for staleness. Dispatched by the code-review-pipeline skill — do not invoke directly.
General code reviewer — bugs, logic, security, and error handling; structural integrity (coupling, cohesion, API surface); and framework best-practices. Dispatched by the code-review-pipeline skill — do not invoke directly.
Reviews test quality, identifies coverage gaps, and flags test antipatterns. Returns list of missing tests. Dispatched by the code-review-pipeline skill — do not invoke directly.
Creates an agent team of parallel dual-engine code reviewers on your git diff, aggregates findings by severity, and fixes critical/high issues. Each reviewer cross-validates with Codex via MCP. Run after implementing a feature or before committing.
Use before implementing any non-trivial feature - validates approaches against real sources using Context7, Serper, GitHub MCPs, and optionally btca for source-level codebase research. Evaluates with dual engines before committing to an implementation
Use when user explicitly requests deep research or comprehensive analysis across many authoritative sources. Creates an agent team for parallel research, gated on answer sufficiency (not a source count), with confidence tracking and structured synthesis. NOT for simple questions answerable with a single search.
Executes bash commands
Hook triggers when Bash tool is used
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
No model invocation
Executes directly as bash, bypassing the AI model
No model invocation
Executes directly as bash, bypassing the AI model
Runs pre-commands
Contains inline bash commands via ! syntax
Runs pre-commands
Contains inline bash commands via ! syntax
Bash prerequisite issue
Uses bash pre-commands but Bash not in allowed tools
Bash prerequisite issue
Uses bash pre-commands but Bash not in allowed tools
Research, planning, and code review skills for Claude Code with dual-engine cross-validation (Claude + Codex).
Pipeline reviewers (3): dispatched by the code-review-pipeline skill based on what the change needs.
Standalone reviewer (1): review-code — reviews completed work against the original plan and coding standards. Invoked via /review-code; carries its own self-contained cross-validation since it runs outside the pipeline.
/research — Start a deep research session/planning — Plan a non-trivial feature before implementation/code-review-pipeline — Run the full review pipeline/review-code — Standalone plan-alignment + standards reviewtask-loop.json has complete: false. Used by research (and available for future long-running skills).The collaboration standard is defined once in the code-review-pipeline skill and injected into every reviewer's task context, so each agent definition stays thin. Following that standard, each reviewer independently:
codex MCP tool with cwd set to the repo rootCross-validated findings (flagged by both engines) receive a confidence boost — these are the highest-signal issues. Reviewers gracefully degrade to Claude-only when Codex is unavailable or returns unusable output.
npm i -g @openai/codexcodex mcp-server — requires Codex CLI installed)bun add -g btca then claude mcp add --transport stdio btca-local -- bunx btca mcp/plugin marketplace add guyathomas/amux-marketplace
/plugin install amux@amux-marketplace
/help
# Should list /amux:research, /amux:code-review-pipeline, etc.
MIT — see LICENSE
npx claudepluginhub guyathomas/workflows --plugin amuxGive Claude Code a second opinion using OpenAI Codex - automatic plan review via hooks
Multi-perspective codebase review with severity-classified findings and holistic remediation
Command-driven Claude/Codex review workflows for multi-round plan and implementation review.
No description provided.
Verification-first engineering toolkit for Claude Code. 15 skills across a 5-phase spine (Investigate → Design → Implement → Verify → Ship), 8 specialist agents, an interactive setup wizard. Every skill has rationalizations + evidence requirements. Built for senior ICs and tech leads.
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.