Updated April 2026

Install Pipelock

Pick your client, copy the command, done.

Run this in a terminal.

Run in terminal

claude mcp add pipelock -- docker pull ghcr.io/luckypipewrench/pipelock:latest

Try It Out

After setup, try these prompts with your AI agent:

→Configure Pipelock to scan all outgoing agent traffic for leaked API keys.

→Set up a forward proxy with Pipelock to intercept and inspect agent HTTP requests.

→Install the community rule bundle to enhance protection against prompt injection.

→Generate a docker-compose configuration to run an agent with full network isolation via Pipelock.

Prerequisites & system requirements

An MCP-compatible client (Claude Code, Cursor, Windsurf, Claude Desktop, or Codex)
Python 3.8+ with pip installed
Docker installed and running

Alternative installation methods

docker pull ghcr.io/luckypipewrench/pipelock:latest

go install github.com/luckyPipewrench/pipelock/cmd/pipelock@latest

Save the docs, env vars, and workflow around Pipelock in Conare so Claude Code, Codex, and Cursor remember it next time.