Factory Ai

What This Skill Does

Factory AI provides a powerful bridge between the OpenClaw agent and the droid CLI tool, enabling automated software engineering workflows directly from your terminal. It acts as an autonomous assistant capable of navigating your codebase, performing complex refactoring, executing build tasks, and managing deployments. By leveraging the Droid engine, this skill allows the agent to maintain context across multi-file projects, interface with MCP (Model Context Protocol) servers for extended tool access, and execute specific git operations, all while maintaining a persistent session for complex, multi-step development tasks.

Installation

The Factory AI skill can be installed via the ClawHub registry. Ensure that the droid binary is correctly installed and accessible in your system path (e.g., at /Users/mitchellbernstein/.local/bin/droid). Once installed, you must authenticate the agent by either running droid login or setting the FACTORY_API_KEY environment variable in your OpenClaw configuration. This key allows the agent to communicate with the underlying Factory AI services, enabling its full suite of coding and deployment capabilities.

Use Cases

Factory AI excels in high-context development scenarios. You can use it to perform deep codebase audits, where the agent analyzes inter-dependent modules to identify logic errors or security vulnerabilities. It is highly effective for automated feature development, where the agent can create new components, write unit tests, and perform refactoring based on natural language instructions. Furthermore, it streamlines the DevOps lifecycle by handling repetitive tasks like writing commit messages, generating pull requests, and initiating deployments to platforms like Fly.io without manual intervention.

Example Prompts

1. "Analyze the current project structure and refactor the authentication middleware in src/middleware/auth.ts to support OAuth2."
2. "Review the pending changes in the current directory and generate a conventional commit message, then push the branch to origin."
3. "Identify potential memory leaks in the React component tree located at @src/components/ and provide a fix using the droid exec command."

Tips & Limitations

For the best results, use the interactive mode when embarking on complex multi-file architectural changes to allow the agent to iterate with you. For single, well-defined tasks, utilize the exec mode with the --json flag if you are parsing the output programmatically within an automation pipeline. Be aware that because this skill has broad access to your filesystem and git history, it is recommended to review all droid exec suggestions before applying them, especially when using the --force flag. Always ensure your environment variables are secured to prevent unauthorized access to your Factory AI API key.