spec-flow
Skill32 starsChanged 9 months agoMITView on GitHub
Spec-driven development workflow. Interactive phase-by-phase confirmation from proposal to implementation. Trigger: 'spec-flow', 'spec mode', 'need a plan', 'structured development', 'write a spec', 'feature spec', 'technical spec', '需求文档', '技术方案', '任务拆解', '规格驱动', '写个方案', '做个规划', '结构化开发', 'plan this feature', 'break this down', 'design doc'. Creates .spec-flow/ directory with proposal, requirements, design, and tasks.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- me-cedric/agent-compasstheir own version, first seen Nov 1, 2026
- plimeor/agent-skillstheir own version, first seen Sep 29, 2026
- SmartAssets-io/skillstheir own version, first seen Aug 12, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- plimeor/agent-skills★ 4same name
- me-cedric/agent-compass★ 3same name
- SmartAssets-io/skills★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
Files
Spec-Flow - Structured Development Workflow
Structured workflow for complex feature development. Creates living documentation that guides implementation and serves as team reference.
Core Principles
- One phase at a time — NEVER generate documents for subsequent phases in advance
- Mandatory confirmation — After each phase, STOP and wait for user confirmation
- User-driven progression — Only proceed when user says "continue"/"ok"/"next"/"继续"/"好"
Parameters
| 参数 | 说明 |
--fast | 跳过逐阶段确认,一次性生成所有文档(最后整体确认) |
--skip-design | 跳过 Design 阶段(简单功能,架构显而易见时) |
Language Rule
所有生成的 .md 文件必须使用中文。
Interaction Rules
Load references/interaction-rules.md for detailed confirmation templates and prohibited behaviors.
Five-Phase Workflow
Copy this checklist and check off items as you complete them:
Spec-Flow Progress:
- Phase 0: Initialize ⚠️ REQUIRED
- 0.1 Run
scripts/init-spec-flow.sh <feature-name>or create.spec-flow/active/<feature>/ - 0.2 Check if
.spec-flow/steering/exists — if so, read for project context - 0.3 问自己:这个功能的核心价值是什么?用一句话能说清吗?
- 0.1 Run
- Phase 1: Proposal ⚠️ REQUIRED
- 1.1 Create
.spec-flow/active/<feature>/proposal.mdusingtemplates/proposal.md.template - 1.2 包含:Background, Goals, Non-Goals, Scope, Risks, Open Questions
- 1.3 问自己:有没有隐含的需求没写出来?
- 1.4 问自己:Non-Goals 是否足够明确?能防止 scope creep 吗?
- 1.5 ⏸️ 确认节点 — 等待用户确认后继续
- 1.1 Create
- Phase 2: Requirements
- 2.1 Create
.spec-flow/active/<feature>/requirements.mdusingtemplates/requirements.md.template - 2.2 Load
references/ears-format.md— 使用 EARS 格式编写需求 - 2.3 包含:FR-xxx 功能需求, NFR-xxx 非功能需求, AC-xxx 验收标准
- 2.4 问自己:每条需求都是可测试的吗?
- 2.5 问自己:边界情况覆盖了吗?错误场景呢?
- 2.6 ⏸️ 确认节点 — 等待用户确认后继续
- 2.1 Create
- Phase 3: Design(除非
--skip-design)- 3.1 Create
.spec-flow/active/<feature>/design.mdusingtemplates/design.md.template - 3.2 包含:Architecture (Mermaid), Components, API, Data Model, Error Handling
- 3.3 问自己:这个设计能满足所有 requirements 吗?逐条对照
- 3.4 问自己:有没有更简单的方案?过度设计了吗?
- 3.5 ⏸️ 确认节点 — 等待用户确认后继续
- 3.1 Create
- Phase 4: Tasks
- 4.1 Create
.spec-flow/active/<feature>/tasks.mdusingtemplates/tasks.md.template - 4.2 Load
references/task-decomposition.md— 遵循任务拆解规范 - 4.3 每个任务:1-2 tool calls 可完成,标注复杂度 Low/Medium/High
- 4.4 分组:Setup → Implementation → Testing → Documentation
- 4.5 问自己:任务之间的依赖关系对吗?有没有可以并行的?
- 4.6 ⏸️ 确认节点 — 等待用户确认后继续
- 4.1 Create
- Phase 5: Implementation
- 5.1 Load
references/execution-modes.md— 确认执行模式 - 5.2 默认 Step Mode(逐任务确认),用户可切换 Batch/Phase Mode
- 5.3 每个任务执行前:读 tasks.md → 检查依赖 → 读 design.md
- 5.4 每个任务执行后:更新 tasks.md 状态(
- [ ]→- [x]) - 5.5 全部完成后:归档到
.spec-flow/archive/
- 5.1 Load
Directory Structure
.spec-flow/
├── steering/ # 全局项目上下文(可选)
│ ├── constitution.md
│ ├── product.md
│ ├── tech.md
│ └── structure.md
├── active/ # 进行中
│ └── <feature>/
│ ├── proposal.md
│ ├── requirements.md
│ ├── design.md
│ └── tasks.md
└── archive/ # 已完成Steering Documents (Optional)
| Document | Purpose | Template |
constitution.md | 项目治理原则 | templates/steering/constitution.md.template |
product.md | 产品愿景、目标用户 | templates/steering/product.md.template |
tech.md | 技术栈、约束 | templates/steering/tech.md.template |
structure.md | 代码组织、命名规范 | templates/steering/structure.md.template |
Phase Transitions
| From | To | Condition |
| Proposal | Requirements | Proposal approved, questions resolved |
| Requirements | Design | Requirements complete, testable |
| Requirements | Tasks | --skip-design, simple feature |
| Design | Tasks | Design approved |
| Tasks | Implementation | Tasks confirmed |
| Implementation | Done | All tasks complete → archive |
Compatibility
Works with any AI agent that supports Skills format: Claude Code, Blade, OpenClaw, Cursor, Windsurf, etc.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

