Claude Gpt Workflow
Claude Code skills for Claude + GPT collaborative workflows. Adversarial plan review + orchestrated code execution with iterative refinement.
Install
npx skills add longranger2/claude-gpt-workflowInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →Claude-GPT Workflow Skills
中文 | English
A collection of Claude Code skills for Claude + GPT automatic collaboration workflows.
Skills
1. codex
Delegates coding tasks to Codex CLI for execution. CodeX is a cost-effective, strong coder — great for batch refactoring, code generation, multi-file changes, and multi-turn implementation tasks.
Based on oil-oil/codex.
Usage:
~/.claude/skills/codex/scripts/ask_codex.sh "Your request"
2. plan-review
Reviews a technical plan via Codex and iteratively refines it. Uses adversarial review to improve plan quality before implementation.
Trigger: /plan-review
3. plan-execute
Executes a finalized plan by delegating coding to Codex. Claude orchestrates, Codex codes, Claude reviews, Codex fixes — iterating until quality passes.
Trigger: /plan-execute
Note
After Claude enters
planmode, plan files are stored under~/.claudeby default. You can change the plan storage directory in~/.claude/settings.json; for example, the config below stores plans in the current project's./plansdirectory:{ "env": {}, "plansDirectory": "./plans" }
Installation
Option 1: npx skills add (Recommended)
npx skills add longranger2/claude-gpt-workflow
Option 2: Manual installation
Copy skills to your Claude Code skills directory:
cp -r plan-review/ ~/.claude/skills/
cp -r plan-execute/ ~/.claude/skills/
cp -r codex/ ~/.claude/skills/
Workflow
%%{init: {'theme': 'base', 'themeVariables': { 'clusterBkg': 'transparent', 'background': 'transparent', 'primaryColor': '#FFF7E6', 'primaryTextColor': '#5A3A00', 'primaryBorderColor': '#F59E0B', 'lineColor': '#6366F1', 'secondaryColor': '#E8F5FF', 'tertiaryColor': '#F3E8FF'}}}%%
flowchart TB
%% ---------- Styles ----------
classDef user fill:#E8F5FF,stroke:#1B6EF3,stroke-width:2px,color:#0B2A5B,stroke-dasharray: 0;
classDef claude fill:#FFF7E6,stroke:#F59E0B,stroke-width:2px,color:#5A3A00,stroke-dasharray: 0;
classDef codex fill:#F3E8FF,stroke:#7C3AED,stroke-width:2px,color:#2E1065,stroke-dasharray: 0;
classDef decision fill:#FFFFFF,stroke:#6366F1,stroke-width:2px,color:#111827,stroke-dasharray: 5 5;
%% ---------- Phase 1: Plan Review ----------
subgraph P1["Phase 1: Plan Review (Adversarial Iteration)"]
direction TB
U1["User: Submit Plan"]:::user
C1["Claude: Delegate to Codex for Critical Review"]:::claude
X1["Codex: Output Review (10+ issues, Critical/High/Medium/Low)"]:::codex
C2["Claude: Evaluate & Refine Plan"]:::claude
D1{"Status?"}:::decision
U1 --> C1 --> X1 --> C2 --> D1
D1 -- "NEEDS_REVISION (Loop A)" --> C1
end
%% ---------- Phase 2: Plan Execute ----------
subgraph P2["Phase 2: Plan Execute (Orchestrated Implementation)"]
direction TB
C3["Claude: Dispatch batch to Codex for implementation"]:::claude
X2["Codex: Implement code changes (Executor)"]:::codex
C4["Claude: Read-only Code Review (Alignment / Quality / Build)"]:::claude
C5["Claude: Write Review (issue list)"]:::claude
D2{"Verdict?"}:::decision
D3{"More steps remaining?"}:::decision
Done([Complete]):::decision
Similar plugs
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local…
Why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.
The 1st Collaborative AI Design Agent Workspace. Local-first and open source. Create prototypes, websites, apps, dashboards, presentations, images, and videos with your team…