SkillAgentSearch skills...

Prospec

Progressive Spec-Driven Development CLI for Brownfield projects. Features Progressive Disclosure (70-80% token savings), Patch Specs, and Multi-Agent support.

Install / Use

/learn @ci-yang/Prospec
About this skill

Quality Score

0/100

Supported Platforms

Universal

README

Prospec

<div align="center">

npm version License: MIT TypeScript Tests Node

Progressive Spec-Driven Development CLI

Empower AI agents with structured workflows for brownfield and greenfield projects

繁體中文Getting StartedDocumentation

</div>

What is Prospec?

Prospec is a CLI tool that bridges the gap between human requirements and AI-driven development. It automates project analysis, knowledge generation, and change management workflows—all while keeping your AI assistant in the loop.

Key Features

  • AI Knowledge Generation — Auto-generate structured knowledge from existing codebases (brownfield) or bootstrap new projects (greenfield)
  • Architecture Analysis — Detect tech stacks, architecture patterns (MVC, Clean Architecture, etc.), and module dependencies
  • AI Agent Agnostic — Works with Claude Code, Gemini CLI, GitHub Copilot, and Codex CLI
  • Progressive Disclosure — Save 70%+ tokens by loading context on-demand
  • Change Management — Structured story → design → plan → tasks workflow with Constitution validation
  • Dual-Layer Specs — Capability specs (living truth) + delta specs (per-change patches) with automatic Spec Sync
  • Skill-Driven — 11 pre-built Skills guide AI through the full SDD lifecycle including UI design, verification, and archiving

Why Prospec?

| Challenge | Solution | |-----------|----------| | AI doesn't know your codebase | prospec knowledge init + /prospec-knowledge-generate auto-scan and generate AI-readable docs | | Context window limitations | Progressive disclosure: load summary first, details on-demand | | Inconsistent AI workflows | Structured Skills enforce story → plan → tasks → implement → verify flow | | Vendor lock-in | Works with 4+ AI CLIs, knowledge stored in universal Markdown | | No design-to-code bridge | /prospec-design generates visual + interaction specs with MCP tool integration | | Knowledge becomes stale | Archive → Knowledge Update feedback loop keeps AI Knowledge in sync |


Installation

# Install as devDependency (recommended)
pnpm add -D prospec

# Or install globally
pnpm add -g prospec

# Verify
prospec --help

Prerequisites


Getting Started

Greenfield Workflow (New Projects)

# 1. Initialize project
mkdir my-project && cd my-project
prospec init --name my-project
# → Select AI assistants (interactive checkbox)
# → Creates .prospec.yaml + directory structure

# 2. Sync AI agent config + generate Skills
prospec agent sync
# → Generates CLAUDE.md + .claude/skills/prospec-*/SKILL.md

# 3. Start developing with Skills (in your AI agent)
/prospec-new-story        # Create change story
/prospec-design           # Generate UI specs (optional)
/prospec-plan             # Generate implementation plan
/prospec-tasks            # Break down tasks
/prospec-implement        # Implement task-by-task
/prospec-verify           # Validate implementation
/prospec-archive          # Archive and sync specs

# Or fast-forward
/prospec-ff               # Generate story → plan → tasks in one go

Brownfield Workflow (Existing Projects)

# 1. Initialize in existing project
cd existing-project
prospec init
# → Auto-detect tech stack
# → Select AI assistants

# 2. Sync AI config + generate Skills
prospec agent sync
# → Generates CLAUDE.md + .claude/skills/prospec-*/SKILL.md

# 3. Scan project and generate raw data
prospec knowledge init
# → Generates raw-scan.md + empty skeleton (_index.md, _conventions.md)

# 4. AI-driven module analysis (in your AI agent)
/prospec-knowledge-generate
# → AI reads raw-scan.md, decides module partitioning
# → Creates modules/*/README.md + fills _index.md

# 5. Develop with Skills
/prospec-explore          # Explore and clarify requirements
/prospec-ff add-feature   # Fast-forward to generate all artifacts
/prospec-implement        # Start coding
/prospec-verify           # Validate against specs
/prospec-archive          # Archive + sync capability specs

CLI Commands

Infrastructure Commands

| Command | Description | |---------|-------------| | prospec init [options] | Initialize Prospec project structure | | prospec knowledge init [--depth <n>] | Scan project and generate raw-scan.md + skeleton | | prospec agent sync [--cli <name>] | Sync AI agent configs + generate Skills |

Change Management Commands

| Command | Description | |---------|-------------| | prospec change story <name> | Create change story (scaffold) | | prospec change plan [--change <name>] | Generate implementation plan (scaffold) | | prospec change tasks [--change <name>] | Break down tasks (scaffold) |

Note: CLI commands create scaffolds; AI agents (via Skills) fill in content.


AI Skills

Prospec generates 11 Skills that guide AI through the full SDD lifecycle:

| Skill | Slash Command | Description | |-------|---------------|-------------| | Explore | /prospec-explore | Think partner for requirement clarification | | New Story | /prospec-new-story | Create structured change story | | Design | /prospec-design | Generate visual + interaction specs (Generate/Extract modes) | | Plan | /prospec-plan | Generate implementation plan + delta-spec | | Tasks | /prospec-tasks | Break down into executable tasks | | Fast-Forward | /prospec-ff | Generate story → plan → tasks in one go | | Implement | /prospec-implement | Implement tasks one-by-one with MCP-first design reading | | Verify | /prospec-verify | 5+1 dimension audit with quality grade (S/A/B/C/D) | | Archive | /prospec-archive | Archive changes + Spec Sync + Knowledge Update prompt | | Knowledge Generate | /prospec-knowledge-generate | AI-driven module analysis and knowledge creation | | Knowledge Update | /prospec-knowledge-update | Incremental knowledge update from delta-spec |

SDD Workflow

Explore → Story → [Design] → Plan → Tasks → Implement → Verify → Archive
 (why)   (what)    (UI)     (how)  (steps)   (code)    (audit)  (wrap up)

Skill Example

# In Claude Code / Gemini CLI / Copilot
/prospec-ff add-authentication

# AI will:
# 1. Call `prospec change story add-authentication`
# 2. Fill in proposal.md (User Story format)
# 3. Call `prospec change plan`
# 4. Fill in plan.md + delta-spec.md
# 5. Call `prospec change tasks`
# 6. Fill in tasks.md (with complexity estimates)
# 7. Output summary + next steps

Architecture

Prospec uses Pragmatic Layered Architecture for CLI development best practices:

src/
├── cli/          — Commander.js commands + formatters
├── services/     — Business logic (10 services)
├── lib/          — Pure utility functions (config, fs, logger, etc.)
├── types/        — Zod schemas + TypeScript types
└── templates/    — Handlebars templates (55 files: 52 .hbs, 3 .md)
    └── skills/   — 11 Skill templates + 22 reference templates

Tech Stack

  • CLI Framework: Commander.js 14 + @inquirer/prompts 8
  • Validation: Zod 4
  • Templating: Handlebars 4.7
  • File Scanning: fast-glob 3.3
  • YAML: eemeli/yaml 2.x (preserves comments)
  • Testing: Vitest 4.0 + memfs
  • TypeScript: 5.9

Testing

# Run all tests (350 tests)
pnpm test

# Watch mode
pnpm run test:watch

# Type check
pnpm run typecheck

# Lint
pnpm run lint

Test Coverage: 350 tests across 4 categories:

  • Unit tests (lib + services): 201 tests
  • Contract tests (CLI output + Skill format): 117 tests
  • Integration tests: 15 tests
  • E2E tests: 17 tests

Project Structure

After running prospec init:

your-project/
├── .prospec.yaml              # Prospec config
├── CLAUDE.md                  # Claude Code config (Layer 0, <100 lines)
├── {base_dir}/
│   ├── CONSTITUTION.md        # Project rules (user-defined)
│   ├── specs/
│   │   ├── capabilities/      # Living behavior specs (accumulated)
│   │   └── history/           # Archived change summaries
│   └── ai-knowledge/
│       ├── _index.md          # Module index (Markdown table)
│       ├── _conventions.md    # Project conventions
│       ├── raw-scan.md        # Auto-generated project scan data
│       ├── module-map.yaml    # Module dependencies
│       └── modules/
│           └── {module}/
│               └── README.md  # Module-specific docs
├── .prospec/                  # Change management (not committed)
│   ├── changes/
│   │   └── {change-name}/
│   │       ├── proposal.md        # User Story + acceptance criteria
│   │       ├── design-spec.md     # Visual spec (optional, UI changes)
│   │       ├── interaction-spec.md # Interaction spec (optional)
│   │       ├── plan.md            # Implementation plan
│   │       ├── tasks.md           # Task breakdown (checkbox format)
│   │       ├── delta-spec.md      # Patch Spec (ADDED/MODIFIED/REMOVED)
│   │       └── metadata.yaml      # Change lifecycle metadata
│   └── archive/               # Archived completed changes
└── .claude/skills/
    ├── prospec-explore/
    ├── prospec-new-story/
    ├── prospec-design/
    ├── prospec-plan/
    ├
View on GitHub
GitHub Stars6
CategoryDevelopment
Updated6d ago
Forks0

Languages

TypeScript

Security Score

85/100

Audited on Mar 30, 2026

No findings