phantom (ghostwright/phantom) is an open-source AI project on GitHub. Repository summary: An AI co-worker with its own computer. Self-evolving, persistent memory, MCP server, secure credential collection, email identity. Built on the Claude Agent SDK. Its focus includes MCP and tool-calling integration, developer-centric engineering workflows, team collaboration integrations. It is suitable for extension, integration, and iterative delivery in real workflows.
License
Apache-2.0
Stars
1,400
Homepage
https://ghostwright.dev/phantomFeatures
- Core capability: An AI co-worker with its own computer. Self-evolving, persistent memory, MCP server, secure credential collection, email identity. Built on the Claude Agent SDK.
- Provides MCP or tool-calling integration
- Built for code generation, debugging, or engineering integration
- Integrates with team collaboration and business systems
- Repository: ghostwright/phantom
- Primary language: TypeScript
Use Cases
- Connects external systems into agent workflows
- Supports AI engineering build-and-iterate workflows for dev teams
- Used for team knowledge collaboration and task follow-ups
- Build internal AI workflow prototypes with phantom
- Validate phantom in production-like engineering scenarios
- Building AI development workflows
FAQ
Teams should first define integration boundaries and call patterns, then map repository capabilities into concrete interfaces, parameters, and access rules. GitHub repository: https://github.com/ghostwright/phantom. Community traction is around 1,400 stars. License: Apache-2.0.
It usually works as an execution component or capability layer, with common deployment fits such as: Connects external systems into agent workflows, Supports AI engineering build-and-iterate workflows for dev teams, Used for team knowledge collaboration and task follow-ups.
