glslang

Plan Execute Judge — Data skill for Claude Code

Data community

Generic plan -> execute -> judge pipeline for the Claude Agent SDK or Codex SDK (TypeScript), with an optional deep-research phase in front.

How to install Plan Execute Judge

This entry records only its repository, not the path inside it, so there is no exact command to give. Open glslang/plan-execute-judge and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Plan Execute Judge does

Generic plan -> execute -> judge pipeline for the Claude Agent SDK or Codex SDK (TypeScript), with an optional deep-research phase in front.

Alternatives in Data

Azure AI Search TypeScript — Azure AI Search for TypeScript 1.8k ★
Agentlas Storm — Run a force-robust Stormbreaker loop — route to real agents, execute a verified pipeline to completion 1.1k ★
Claude Deep Research Skill — Enterprise-grade deep research skill for Claude Code with 8-phase pipeline, source credibility scoring, and au 1k ★

README

plan-execute-judge

[![CI](https://github.com/glslang/plan-execute-judge/actions/workflows/ci.yml/badge.svg)](https://github.com/glslang/plan-execute-judge/actions/workflows/ci.yml) [![Node.js](https://img.shields.io/badge/node-%3E%3D22.9-339933?logo=node.js&logoColor=white)](https://nodejs.org) [![TypeScript](https://img.shields.io/badge/TypeScript-5.5-3178C6?logo=typescript&logoColor=white)](https://www.typescriptlang.org) [![Claude Agent SDK](https://img.shields.io/badge/Claude%20Agent%20SDK-0.3-D97757?logo=anthropic&logoColor=white)](https://www.npmjs.com/package/@anthropic-ai/claude-agent-sdk) [![Codex SDK](https://img.shields.io/badge/Codex%20SDK-0.144-111827?logo=openai&logoColor=white)](https://www.npmjs.com/package/@openai/codex-sdk) [![License: MIT](https://img.shields.io/badge/License-MIT-yellow)](LICENSE)

Generic plan -> execute -> judge pipeline for the Claude Agent SDK or Codex SDK (TypeScript), with an optional deep-research phase in front. Research and planning can fan out across multiple independent agents; when multiple plans are produced, a `refinements` phase merges them into one final plan. Fresh agent runs with real control flow. The research brief, the plan text, and a typed verdict are the only data crossing phase boundaries -- never a conversation transcript. `RESEARCH.md` and `PLAN.md` are written as human-readable artifacts of the run.

task ─┬─> [research, optional] ingests PDFs, web pages, git repos + your own
      │       notes ─> research brief (saved to RESEARCH.md)
      │                     │
      v                     v
      plan (read-only research) ─> candidate plan(s)
                                          │
            ┌─────────────────────────────┘
            v
      [refinements, when needed] merge candidate plans -> PLAN.md
            |
            v
      [approval, optional] user verifies/agrees to PLAN.md
            |
            v
      execute (acceptEdits -- implements the plan, leaves changes uncommi