Pilot Shell

v1.0.0Coding Agentsstable

How real engineers run Claude Code: spec-driven planning, enforced TDD, persistent memory, and quality enforcement on all levels. Make Claude Code production-ready.

ai-agentsai-assistantai-codingai-coding-toolsai-engineering
Share:
1,725
Stars
0
Downloads
0
Weekly
0/5

What is Pilot Shell?

Pilot Shell is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to how real engineers run claude code: spec-driven planning, enforced tdd, persistent memory, and quality enforcement on all levels. make claude code production-ready.

How real engineers run Claude Code: spec-driven planning, enforced TDD, persistent memory, and quality enforcement on all levels. Make Claude Code production-ready.

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • How real engineers run Claude Code: spec-driven planning, en

Use Cases

Production-ready Claude Code with spec-driven planning
TDD enforcement and quality validation
maxritter

Maintainer

LicenseNOASSERTION
Languagetypescript
Versionv1.0.0
UpdatedMay 21, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx pilot-shell

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Pilot Shell

Pilot Shell is a framework that enhances Claude Code and Codex CLI with production-grade software engineering workflows including spec-driven planning, enforced test-driven development, quality gates (linting, formatting, type checking), and a persistent knowledge system that preserves architectural decisions across sessions. It installs globally as a shell script and adds slash commands (/prd, /spec, /fix, /setup-rules) that guide AI agents through structured development cycles rather than ad-hoc code generation. Engineering teams use Pilot Shell to make AI-generated code reliable enough for production by ensuring every feature goes through requirements, implementation, and verification phases with measurable quality checkpoints.

Prerequisites

  • macOS, Linux, or Windows with WSL2
  • Claude Code installed with an active Anthropic Claude Max (5x/20x or higher) subscription, OR Codex CLI with an OpenAI Plus/Pro subscription
  • bash or zsh shell (WSL2 on Windows)
  • curl available for the install script
  • Git repository for the projects you want to work with
1

Install Pilot Shell globally

Run the one-line install script. It installs Pilot Shell globally in under 2 minutes and is available for macOS, Linux, and Windows WSL2.

curl -fsSL https://raw.githubusercontent.com/maxritter/pilot-shell/main/install.sh | bash
2

Set up project-specific rules

Navigate to your project directory and start Claude Code. Run the /setup-rules command so Pilot Shell discovers your codebase conventions (testing framework, linting config, code style) and generates modular rule files.

cd your-project
claude
# Then in the Claude session:
/setup-rules
3

Create a product requirements document

Use the /prd command to turn a vague idea into a structured requirements document. Pilot Shell guides the conversation through scope, acceptance criteria, and optional research tiers.

# In your Claude Code session:
/prd
4

Implement a feature with spec-driven workflow

Use the /spec command to implement a feature. Pilot Shell creates a specification, implements it with TDD, enforces quality gates, and runs verification—all within a single structured session.

# In your Claude Code session:
/spec
5

Fix bugs with the structured bugfix workflow

Use the /fix command for bug reports. Pilot Shell investigates the root cause, writes a failing test, implements the fix, and verifies it passes all quality gates.

# In your Claude Code session:
/fix
6

Monitor progress via the local console

Pilot Shell runs a local web dashboard at http://localhost:41777 that shows real-time session progress, specifications, knowledge graph state, and token usage across all active Claude Code sessions.

# Access in browser:
open http://localhost:41777

Pilot Shell Examples

Claude Code MCP integration (Pilot Shell as MCP server)

Pilot Shell works as a framework on top of Claude Code rather than as a standalone MCP server. A typical claude_desktop_config.json for a project using Pilot Shell looks like this.

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/your/project"]
    }
  }
}

Prompts to try

Use these slash commands inside Claude Code after installing Pilot Shell and running /setup-rules in your project.

- "/prd — Start a requirements brainstorming session for a new feature"
- "/spec — Implement the feature from the PRD with TDD and quality gates"
- "/fix — Investigate and fix a reported bug with test coverage"
- "/benchmark — Measure the impact of your rule and skill changes"
- "/create-skill — Extract domain knowledge into a reusable skill"

Troubleshooting Pilot Shell

Install script fails with a permission denied error

Ensure curl is installed and you have write access to /usr/local/bin or your ~/.local/bin directory. On macOS you may need to grant Terminal full disk access in System Settings → Privacy & Security. Re-run with: curl -fsSL https://raw.githubusercontent.com/maxritter/pilot-shell/main/install.sh | bash.

/setup-rules does not detect the project's test framework

Ensure you are running Claude Code from the project root directory where package.json, pyproject.toml, or equivalent config files exist. Pilot Shell scans the current directory for framework signals; running it from a parent directory will miss project-specific config.

Quality gates fail during /spec with linting errors

This is intentional—Pilot Shell enforces quality gates before declaring a feature complete. Fix the reported lint or type errors, then allow Pilot Shell to re-run the verification step. You can also run /setup-rules again to update the rules if your linting config has changed since initial setup.

Frequently Asked Questions about Pilot Shell

What is Pilot Shell?

Pilot Shell is a Model Context Protocol (MCP) server that how real engineers run claude code: spec-driven planning, enforced tdd, persistent memory, and quality enforcement on all levels. make claude code production-ready. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Pilot Shell?

Follow the installation instructions on the Pilot Shell GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Pilot Shell?

Pilot Shell works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Pilot Shell free to use?

Yes, Pilot Shell is open source and available under the NOASSERTION license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "pilot-shell": { "command": "npx", "args": ["-y", "pilot-shell"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Pilot Shell?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides