Automate pre-development workflows for large-scale project rewrites, migrations, refactors, and overhauls. Analyze full codebase architecture and dependencies, decompose into phased task plans with Mermaid diagrams, effort estimates, and acceptance criteria; generate structured docs, track progress via MASTER.md, create sub-SKILLS, and execute isolated tasks with tests.
Performs deep codebase analysis for the Spec-Driven Develop workflow. Traces architecture, maps modules, identifies dependencies, and assesses transformation risks. Returns structured analysis data for document generation.
Designs phased task decomposition for large-scale project transformations. Takes analysis data and target state as input, produces a dependency-aware implementation plan with milestones, effort estimates, and acceptance criteria.
Executes a single development task from the phased plan. Receives task description, acceptance criteria, relevant file paths, and coding standards. Implements the change, writes/runs tests, and returns a structured completion report. Designed for parallel execution — multiple instances work on independent tasks simultaneously using isolated worktrees.
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.
English | 中文
One Markdown file. Any coding agent. Full pre-development automation.
Spec-Driven Develop is a platform-agnostic AI agent skill that automates the pre-development workflow for large-scale complex tasks. It is not a framework, not a runtime, not a package manager — it is a single SKILL.md file that teaches any AI coding agent a structured methodology.
The core mechanism is dead simple: your agent reads a Markdown file containing structured instructions, then follows them. No SDK, no API integration, no platform-specific hooks. If your coding agent can read Markdown — and they all can — it works.
When you tell your agent something like "rewrite this project in Rust" or "migrate to a microservice architecture", it automatically kicks off a standardized preparation pipeline before writing a single line of code:
A master progress file (docs/progress/MASTER.md) serves as the agent's "memory anchor" across conversations, so it never loses track of where things stand — no matter how many sessions it takes.
The Claude Code ecosystem now has full-blown frameworks with dozens of agents, multi-phase pipelines, and opinionated workflows. They're powerful — but they're also heavy, and they lock you into a single platform.
| Spec-Driven Develop | Superpowers | oh-my-claudecode | |
|---|---|---|---|
| What it is | A single SKILL file | Full skills framework + methodology | Multi-agent orchestration system |
| Core files | 1 Markdown file (~200 lines) | Plugin with multiple skills, agents, hooks | Plugin with 32+ specialized agents |
| Dependencies | None | Requires Claude Code plugin system | Requires Claude Code plugin system |
| Methodology | Document-driven planning | Enforced TDD (RED-GREEN-REFACTOR) | Team-based multi-agent delegation |
| Cross-platform | Any agent that reads Markdown | Claude Code (primary) | Claude Code only |
| Philosophy | Do one thing, do it well | Complete development methodology | Parallel multi-agent orchestration |
Spec-Driven Develop takes a fundamentally different approach: instead of wrapping your agent in a framework, it gives the agent a methodology through a plain Markdown file. No hooks, no runtime overhead, no forced workflows. You keep full control.
Lightweight does not mean weak. A well-structured Markdown file can carry a surprisingly sophisticated workflow — project analysis, phased task decomposition, progress tracking, sub-SKILL generation — all without a single line of executable code. The agent reads the instructions and executes them. That's it. The simplicity is the feature.
This makes it especially suited for:
The SKILL prompt is written in a generic, platform-neutral way. It gracefully degrades on platforms without certain capabilities — for example, if sub-agents aren't available, it falls back to sequential execution automatically.
Tested platforms with install scripts:
Any other agent — copy SKILL.md to wherever your agent reads instructions. That's it. The file has no external dependencies and no platform-specific logic. It works with Windsurf, Cline, Aider, Continue, Roo Code, Augment, or any other coding agent that reads Markdown-based skills or system prompts.
/plugin marketplace add zhu1090093659/spec_driven_develop
/plugin install spec-driven-develop@spec-driven-develop
After installation, run /reload-plugins to activate.
Use the built-in skill installer (inside a Codex session):
$skill-installer install https://github.com/zhu1090093659/spec_driven_develop/tree/main/plugins/spec-driven-develop/skills/spec-driven-develop
Or install via shell:
bash <(curl -sL https://raw.githubusercontent.com/zhu1090093659/spec_driven_develop/main/scripts/install-codex.sh)
bash <(curl -sL https://raw.githubusercontent.com/zhu1090093659/spec_driven_develop/main/scripts/install-cursor.sh)
Or clone the repo and run locally:
git clone https://github.com/zhu1090093659/spec_driven_develop.git
bash spec_driven_develop/scripts/install-cursor.sh
四个思考训练 skill:taste-audit、intent-refine、judgment-redteam、abstraction-uplift。AI 时代最反潮流的一套 skill——不替你思考,只逼你思考。
npx claudepluginhub zhu1090093659/spec_driven_develop --plugin spec-driven-develop11 agents, 35 skills, 18 commands, 9 hooks — spec-driven multi-agent orchestration for Claude Code, with optional cross-device semantic memory.
Plan and autonomously build a software task end-to-end. Recons the codebase, applies preloaded memory, decomposes into the right number of phases, gets one confirmation, then prepares a single ready-to-paste /goal command — one paste between you and done — that drives execution to completion with built-in retry, fix-spec recovery, and per-phase memory writeback. Works on Claude Code and Codex.
Spec-driven development plugin for Claude Code. Markdown specs as the source of truth, code downstream.
AI-powered cascading development framework with design document system and multi-agent collaboration. Breaks down projects into Features (Mega Plan), Features into Stories (Hybrid Ralph), with auto-generated technical design docs, dependency-driven batch execution, Git Worktree isolation, and support for multiple AI agents (Codex, Amp, Aider, etc.).
GSD Core is a meta-prompting, context engineering, and spec-driven development system for AI coding agents.
Production-grade engineering skills for AI coding agents — covering the full software development lifecycle from spec to ship.