spec-server 0.6.2

MCP server for structured feature development through Requirements → Design → Tasks workflow

spec-server

An MCP (Model Context Protocol) server that provides structured feature development capabilities through a systematic three-phase workflow: Requirements → Design → Implementation Tasks.

Overview

spec-server helps AI assistants and developers transform rough feature ideas into executable implementation plans through a structured approach:

1. Requirements Phase: Define user stories and acceptance criteria in EARS format
2. Design Phase: Create comprehensive technical design documents
3. Tasks Phase: Generate actionable implementation tasks with test-driven development focus

Features

• Systematic Workflow: Enforced progression through Requirements → Design → Tasks phases
• Enhanced Design Format: Intent/Goals/Logic structure for all technical elements
• Auto-Format Task Lists: Automatic formatting and requirements linking for tasks
• LLM Guidance: Built-in guidance system for AI assistants with conversation templates
• MCP Integration: Works with any MCP-compatible AI assistant
• File References: Support for #[[file:path]] syntax to include external specifications
• Task Management: Hierarchical task tracking with status updates and completion validation
• Multiple Transports: Support for both stdio and SSE (Server-Sent Events) transport methods
• Validation: Built-in validation for document formats and workflow transitions

Installation

From PyPI

pip install spec-server

From Source

git clone https://github.com/teknologika/spec-server.git
cd spec-server
pip install -e .

Usage

As MCP Server

Add to your MCP client configuration:

{
  "mcpServers": {
    "spec-server": {
      "command": "spec-server",
      "args": ["stdio"],
      "disabled": false
    }
  }
}

Command Line

# Run with stdio transport (default)
spec-server

# Run with SSE transport on port 8000
spec-server sse 8000

# Run with SSE transport on custom port
spec-server sse 3000

MCP Tools

The server exposes the following MCP tools:

create_spec

Create a new feature specification with initial requirements document.

Parameters:

• feature_name (string): Kebab-case feature identifier
• initial_idea (string): User's rough feature description

update_spec_document

Update requirements, design, or tasks documents with workflow validation.

Parameters:

• feature_name (string): Target spec identifier
• document_type (enum): "requirements" | "design" | "tasks"
• content (string): Updated document content
• phase_approval (boolean): Whether user approves current phase

list_specs

List all existing specifications with their current status and progress.

read_spec_document

Retrieve content of spec documents with file reference resolution.

Parameters:

• feature_name (string): Target spec identifier
• document_type (enum): "requirements" | "design" | "tasks"

execute_task

Execute a specific implementation task from the tasks document.

Parameters:

• feature_name (string): Target spec identifier
• task_identifier (string): Task number/identifier

delete_spec

Remove a specification entirely including all documents.

Parameters:

• feature_name (string): Target spec identifier

get_full_guidance

Get the complete LLM guidance document for effective spec-server usage.

Returns: Complete guidance on conversational approach, enhanced design format, and auto-format task lists.

get_guidance

Get phase-specific guidance for requirements, design, or tasks phases.

Parameters:

• phase (string): "requirements" | "design" | "tasks" | "general"

Returns: Phase-specific conversation starters, templates, and best practices.

LLM Guidance System

spec-server includes a comprehensive guidance system designed to help AI assistants use the tool effectively through conversational interactions with users.

Accessing Guidance

For AI Assistants:

• Use get_full_guidance() to access the complete guidance document
• Use get_guidance(phase="requirements|design|tasks") for phase-specific guidance
• Guidance includes conversation starters, templates, and best practices

For Developers:

• Guidance is available at docs/llm-guidance.md
• Automatically accessible through MCP tools when using spec-server
• Always up-to-date with the installed version

Enhanced Features

Enhanced Design Format:

• All technical elements use Intent/Goals/Logic structure
• Intent: Brief description of purpose
• Goals: Specific, measurable objectives
• Logic: Detailed implementation approach

Auto-Format Task Lists:

• Tasks are automatically formatted into consistent structure
• Requirements references are automatically added
• Non-task content is moved to appropriate documents
• Task completion is validated against requirements and design

Workflow

1. Requirements Phase

• Create user stories in "As a [role], I want [feature], so that [benefit]" format
• Define acceptance criteria using EARS (Easy Approach to Requirements Syntax)
• Must receive explicit approval before advancing to design phase

2. Design Phase

• Generate comprehensive technical design based on approved requirements
• Use Intent/Goals/Logic structure for all technical elements
• Include sections: Overview, Architecture, Components, Data Models, Error Handling, Testing
• Conduct research and incorporate findings into design decisions
• Must receive explicit approval before advancing to tasks phase

3. Tasks Phase

• Create actionable implementation tasks focused on code development
• System automatically formats tasks and links to requirements
• Tasks are validated against requirements and design upon completion
• Tasks ready for execution by coding agents

File Structure

specs/
├── feature-name-1/
│   ├── requirements.md          # EARS format requirements
│   ├── design.md               # Intent/Goals/Logic format design
│   └── tasks.md                # Auto-formatted tasks with requirements links
├── feature-name-2/
│   ├── requirements.md
│   └── design.md
├── completed/                   # Automatically moved completed specs
│   └── old-feature/
│       ├── requirements.md
│       ├── design.md
│       └── tasks.md
└── .spec-metadata.json         # Spec tracking and progress metadata

File References

Spec documents support file references using the syntax #[[file:relative/path/to/file.md]]. These references are automatically resolved and their content is included when documents are read.

Example:

# API Design

The API follows the OpenAPI specification defined in:
#[[file:api/openapi.yaml]]

Development

Setup Development Environment

git clone https://github.com/teknologika/spec-server.git
cd spec-server
pip install -e ".[dev]"

Run Tests

pytest

Code Quality

# Format code
black src tests

# Sort imports
isort src tests

# Lint code
flake8 src tests

# Type checking
mypy src

Configuration

Workspace Integration

spec-server features intelligent workspace detection that automatically organizes your specifications with your project files when running in an IDE or project directory.

How It Works:

• Automatic Detection: Scans upward from current directory looking for workspace indicators
• Smart Placement: Creates .specs/ directory at the detected workspace root
• Fallback Behavior: Uses specs/ directory in current location if no workspace detected
• IDE Friendly: Works seamlessly with VS Code, IntelliJ, Sublime Text, and other editors

Workspace Indicators:

• .git (Git repository)
• package.json (Node.js project)
• pyproject.toml (Python project)
• Cargo.toml (Rust project)
• go.mod (Go project)
• pom.xml (Maven project)
• .vscode, .idea (IDE configurations)
• README.md, LICENSE, Makefile (common project files)

Example Structure:

my-project/                 ← Detected workspace root
├── .git/
├── src/
├── package.json
├── README.md
└── .specs/                 ← Specs automatically placed here
    ├── user-auth/
    │   ├── requirements.md
    │   ├── design.md
    │   └── tasks.md
    ├── data-export/
    │   └── requirements.md
    └── .spec-metadata.json

Benefits:

• Version Control: Specs can be committed alongside your code
• Team Collaboration: Shared specifications across team members
• Context Awareness: Specs are logically grouped with related projects
• IDE Integration: Specifications appear in your project file tree
• Automatic Organization: No manual directory management required

Configuration:

# Enable/disable workspace detection (default: true)
export SPEC_SERVER_AUTO_DETECT_WORKSPACE=true

# Customize specs directory name in workspace (default: ".specs")
export SPEC_SERVER_WORKSPACE_SPECS_DIR=".my-specs"

# Customize fallback directory name (default: "specs")
export SPEC_SERVER_SPECS_DIR="my-specs"

Environment Variables

• SPEC_SERVER_SPECS_DIR: Base directory for specs (default: "specs")
• SPEC_SERVER_AUTO_DETECT_WORKSPACE: Enable workspace detection (default: "true")
• SPEC_SERVER_WORKSPACE_SPECS_DIR: Specs directory name in workspace (default: ".specs")
• SPEC_SERVER_PORT: Default SSE server port (default: 8000)
• SPEC_SERVER_LOG_LEVEL: Logging level (default: "INFO")

Configuration File

Optional spec-server.json:

{
  "specs_dir": "specs",
  "auto_detect_workspace": true,
  "workspace_specs_dir": ".specs",
  "host": "127.0.0.1",
  "port": 8000,
  "transport": "stdio",
  "log_level": "INFO",
  "auto_backup": true,
  "cache_enabled": true
}

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Support

• GitHub Issues: https://github.com/teknologika/spec-server/issues
• LLM Guidance: docs/llm-guidance.md
• Documentation: [Coming Soon]

Roadmap