A CLI tool that generates comprehensive documentation for AI coding agents (Claude Code, Gemini CLI, OpenCode, etc.).
Proact creates documentation that instructs AI coding agents to:
- Follow best development practices and processes
- Apply continuous improvement based on feedback
- Understand how to install and use tools like Playwright MCP
- Maintain code quality through testing, linting, and documentation
# Basic usage
cargo run -- <target-project-path>
# With verbose output (shows file operations)
cargo run -- -v <target-project-path>
# Dry run mode (preview without creating files)
cargo run -- -n <target-project-path>
# Specify custom output directory
cargo run -- -o ./custom-docs <target-project-path>
# Example: Generate docs for a neighboring project
cargo run -- -v ../my-project
# Example: Preview what would be created
cargo run -- --dry-run ../my-project-V, --version- Show version information-h, --help- Display help message-v, --verbose- Enable verbose output showing generation progress and file operations-n, --dry-run- Show what would be done without actually creating files (implies verbose)-o, --output-dir <DIR>- Output directory for generated documentation (default:./docs)<TARGET>- Required: Path to an existing project directory
The tool generates an ai_agent_instructions.md file containing:
- Process Guidelines - Checkpoint-based development workflow
- Quality Standards - Documentation, testing, and code quality requirements
- Continuous Improvement - Learning from failures and updating processes
- Playwright MCP Setup - Installation and usage instructions for browser automation
- Project-Specific Notes - Automatically detects project type and adds relevant commands
Proact automatically detects the project type and includes appropriate commands:
- Rust - cargo commands for build, test, clippy, fmt
- JavaScript/Node.js - npm/yarn commands for dependencies, testing, linting
- Python - pip/poetry, pytest, black, pylint commands
- Go - go build, test, vet, fmt commands
# Build the project
cargo build
# Run tests
cargo test
# Run linting
cargo clippy --all-targets --all-features -- -D warnings
# Format code
cargo fmt
# Build release version
cargo build --releaseFuture versions will include:
- Additional MCP server documentation beyond Playwright
- More language-specific guidelines
- Customizable templates
- Configuration file support
MIT