</>MCP Agents Market
Skill

spec_driven_develop

by zhu1090093659979ShellUpdated 2026-07-26

Spec-driven development workflow for AI coding agents: architecture-first planning, task decomposition, GitHub Issue/PR tracking, Deep Discuss, and adaptive control for Claude Code, Codex, Cursor, and other Markdown-capable agents.

Claude CodeCodexOpenCodeCursorWindsurfClineAiderContinue

Spec-Driven Develop is an architecture-first workflow agent skill that brings spec-driven development methodology to AI coding agents like Claude Code, Codex, Cursor, and Windsurf. It transforms large software changes into a systematic seven-phase pipeline—from deep project analysis and task decomposition to GitHub Issue/PR tracking and adaptive control feedback loops. The skill includes three complementary workflows: Spec-Driven Develop for complex implementation projects, Deep Discuss for structured technical analysis, and Review SPD for findings-first code reviews. Built entirely in Markdown with no runtime dependencies, it enforces S.U.P.E.R. architecture principles (Single Purpose, Unidirectional Flow, Ports over Implementation, Environment-Agnostic, Replaceable Parts) throughout the development lifecycle.

Key Features

Seven-phase development pipeline: intent capture, deep analysis with S.U.P.E.R. health evaluation, scope refinement, task decomposition, progress tracking, execution with adaptive control, and artifact archiving
Automatic GitHub integration with Issues, Milestones, Labels, Projects boards, worktrees, and coherent batch PRs that group related tasks for reviewable integration
Adaptive control system inspired by engineering cybernetics that monitors execution drift and automatically corrects course when plans diverge from reality (20%/40%/60% thresholds)
Deep Discuss workflow for structured problem analysis through seven phases of listening, validation, root-cause analysis, solution design, and self-review before execution
Review SPD findings-first code review that prioritizes bugs, regressions, and correctness issues across uncommitted changes, date-range commits, or branch/PR diffs
S.U.P.E.R. architecture principles enforced at every phase with per-module compliance scoring, violation hotspots, and per-task design driver annotations
Orchestrator-centric execution with tiered sub-agent dispatch (Tier 0 direct, Tier 1 task-executor, Tier 2 parallel lanes) and tiered review (L1 validation, L2 diff review, L3 independent reviewer)
Platform-agnostic Markdown design with graceful degradation and native plugin support for Claude Code, Codex, and OpenCode

Use Cases

  • 01Migrating a legacy codebase to a modern language or framework with structured task tracking and architecture compliance checks
  • 02Refactoring a monolith into microservices with dependency analysis, parallel execution lanes, and S.U.P.E.R. principle enforcement
  • 03Conducting structured technical discussions to analyze performance bottlenecks, evaluate architecture trade-offs, or design solutions before implementation
  • 04Performing findings-first code reviews focused on bugs and regressions for uncommitted changes, recent commits, or pull requests
  • 05Managing large-scale feature development with automatic GitHub Issue creation, milestone organization, and batch PR integration
  • 06Resuming interrupted development sessions by reading MASTER.md progress files and GitHub Issue states to restore context across conversations

Related Skills

View more

superpowers

An agentic skills framework & software development methodology that works.

Skill294k

skills

Skills for Real Engineers. Straight from my .agents directory.

Skill273.9k

skills

Public repository for Agent Skills

Skill179.3k

ponytail

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

Skill150.6k

spec_driven_develop — FAQ

What is Spec-Driven Develop?+

Spec-Driven Develop is a pure-Markdown spec-driven development workflow for AI coding agents that provides a repeatable system for analyzing codebases, refining scope, breaking work into tasks, tracking progress, executing changes, and archiving results with traceable artifacts. It includes three complementary workflows: Spec-Driven Develop for large-scale implementation, Deep Discuss for structured analysis, and Review SPD for code review.

How do I install Spec-Driven Develop in Claude Code?+

Run `/plugin marketplace add zhu1090093659/spec_driven_develop` followed by `/plugin install spec-driven-develop@spec-driven-develop`, then run `/reload-plugins` to activate. For Codex, use `codex plugin marketplace add zhu1090093659/spec_driven_develop --ref main` or the bash installer script. For Cursor and other agents, run the platform-specific install script or manually copy SKILL.md to your agent's skills directory.

Which AI coding agents work with this skill?+

It works with Claude Code, Codex, OpenCode, Cursor, Windsurf, Cline, Aider, Continue, Roo Code, Augment, and any other Markdown-capable coding agent that reads custom skills or system prompts. The skill includes native plugin support for Claude Code, Codex, and OpenCode, plus installation scripts for Cursor.

Does Spec-Driven Develop require API keys or external dependencies?+

No runtime dependencies or API keys are required for the core workflow, which is pure Markdown. GitHub-native tracking uses the `gh` CLI when available but gracefully degrades to local-only Markdown mode if unavailable or if the repo isn't on GitHub.

Is Spec-Driven Develop free to use?+

Yes, it is open source under the MIT license and completely free to use, modify, and distribute.

When should I use Deep Discuss versus Spec-Driven Develop?+

Use Deep Discuss when the task is ambiguous and requires technical analysis, debugging direction, architecture choices, or brainstorming. Use Spec-Driven Develop when the implementation goal is clear enough to plan and execute as a structured multi-phase project with task decomposition and progress tracking.

How do I install spec_driven_develop?+

Open the source repository on GitHub and follow its README. spec_driven_develop is a skill — MCP Agents Market links you directly to the official repo.

Is spec_driven_develop free?+

spec_driven_develop is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches

Developer Tools SkillsPlanning SkillsCode Review SkillsBest MCP ServersBest PluginsBest AgentsTop rated listingsClaude Code toolsBrowse all categories