Claude Code Runner

Claude Code Runner

Overview

A wrapper skill for running Claude Code programmatically in non-interactive environments. Uses PTY (pseudo-terminal) to handle TTY-required operations and automatically responds to confirmation prompts.

Features

Installation

# Clone the skill
git clone https://github.com/lhl09120/claude-code-runner-en.git
Make script executable
chmod +x claude-code-runner-en/scripts/run_claude.py

Usage

Basic Usage

from claude_code_runner import run_claude_code
result = run_claude_code(
    workdir='/path/to/project',
    prompt='Refactor the authentication module to use JWT tokens',
    user='lighthouse',
    timeout=300
)
print(result)

Via Command Line

python3 scripts/run_claude.py /path/to/project "Your task description here"

Advanced Options

result = run_claude_code(
    workdir='/root/repo/my-project',
    prompt='''
    1. Review the codebase
    2. Identify security vulnerabilities
    3. Fix any issues found
    4. Add appropriate tests
    ''',
    user='developer',
    timeout=600  # 10 minutes
)

API Reference

run_claude_code(workdir, prompt, user='lighthouse', timeout=300)

Execute a Claude Code task in a PTY environment.

Parameters:

Returns:

Behavior: 1. Copies project to temporary directory 2. Changes ownership to specified user 3. Executes Claude Code via PTY 4. Auto-responds to confirmation prompts 5. Syncs changes back to original directory 6. Cleans up temporary files

Use Cases

1. Automated Code Review

result = run_claude_code(
    workdir='/root/repo/project',
    prompt='Review this codebase and identify potential bugs or improvements'
)

2. Refactoring Tasks

result = run_claude_code(
    workdir='/root/repo/legacy-app',
    prompt='Refactor the database layer to use SQLAlchemy ORM instead of raw SQL'
)

3. Adding Features

result = run_claude_code(
    workdir='/root/repo/api-service',
    prompt='''
    Add a new REST endpoint for user profile management:
    - GET /api/users/{id}/profile
    - PUT /api/users/{id}/profile
    - Include validation and error handling
    - Add unit tests
    '''
)

4. Bug Fixes

result = run_claude_code(
    workdir='/root/repo/web-app',
    prompt='Fix the memory leak in the WebSocket connection handler'
)

Requirements

Configuration

Environment Variables

Customization

Edit scripts/run_claude.py to customize:

Troubleshooting

"Permission denied" errors

Ensure the script is run with sufficient privileges to:

Claude Code not found

Make sure Claude Code is installed and in the system PATH:

which claude

Task timeout

Increase the timeout for long-running tasks:

run_claude_code(workdir, prompt, timeout=600)  # 10 minutes

Interactive prompts not auto-responded

Add new prompt patterns to the auto-respond logic:

if b'new prompt text' in output:
    os.write(master_fd, b'y\n')

Limitations

License

MIT License

Copyright (c) 2026 lhl09120

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

See LICENSE file for full details.

Changelog

v1.0.0 (2026-02-27)

Basic Usage

from claude_code_runner import run_claude_code
result = run_claude_code(
    workdir='/path/to/project',
    prompt='Refactor the authentication module to use JWT tokens',
    user='lighthouse',
    timeout=300
)
print(result)

Via Command Line

python3 scripts/run_claude.py /path/to/project "Your task description here"

Advanced Options

result = run_claude_code(
    workdir='/root/repo/my-project',
    prompt='''
    1. Review the codebase
    2. Identify security vulnerabilities
    3. Fix any issues found
    4. Add appropriate tests
    ''',
    user='developer',
    timeout=600  # 10 minutes
)

Environment Variables

Customization

Edit scripts/run_claude.py to customize:

"Permission denied" errors

Ensure the script is run with sufficient privileges to:

Claude Code not found

Make sure Claude Code is installed and in the system PATH:

which claude

Task timeout

Increase the timeout for long-running tasks:

run_claude_code(workdir, prompt, timeout=600)  # 10 minutes

Interactive prompts not auto-responded

Add new prompt patterns to the auto-respond logic:

if b'new prompt text' in output:
    os.write(master_fd, b'y\n')