Resource Recommendation
Display Name: cowork-to-code-bridge
Category: Tooling
Sub-Category: Developer Tools / MCP Servers / Agent Infrastructure
Primary Link: https://github.com/abhinaykrupa/cowork-to-code-bridge
Author Name: Abhinay Krupa
Author Link: https://github.com/abhinaykrupa
License: MIT
Description
An async file-based bridge that lets Claude Cowork (and any MCP-aware agent like CrewAI, AutoGen, LlamaIndex) hand tasks to a local Claude Code agent on your Mac/Linux machine. The agent gets full machine context — repos, shell, environment, installed tools — with no open ports, HMAC token auth, and idempotent retry semantics. Ships with 15+ starter scripts, a selfcheck command, production safety features (per-op metrics, loop detection, budget caps), and installs in one command via Homebrew or curl.
Why it's valuable to Claude Code users
This is the missing piece for Claude Cowork users who want Claude Code's full agent capabilities (file editing, shell execution, git, deploys) from within a Cowork chat. It also serves as an MCP server that any Claude Code-aware agent framework can call to escalate complex tasks to a local Claude Code instance — addressing a gap where sandbox-based agents can't access the real machine context Claude Code needs.
Resource Recommendation
Display Name: cowork-to-code-bridge
Category: Tooling
Sub-Category: Developer Tools / MCP Servers / Agent Infrastructure
Primary Link: https://github.com/abhinaykrupa/cowork-to-code-bridge
Author Name: Abhinay Krupa
Author Link: https://github.com/abhinaykrupa
License: MIT
Description
An async file-based bridge that lets Claude Cowork (and any MCP-aware agent like CrewAI, AutoGen, LlamaIndex) hand tasks to a local Claude Code agent on your Mac/Linux machine. The agent gets full machine context — repos, shell, environment, installed tools — with no open ports, HMAC token auth, and idempotent retry semantics. Ships with 15+ starter scripts, a selfcheck command, production safety features (per-op metrics, loop detection, budget caps), and installs in one command via Homebrew or curl.
Why it's valuable to Claude Code users
This is the missing piece for Claude Cowork users who want Claude Code's full agent capabilities (file editing, shell execution, git, deploys) from within a Cowork chat. It also serves as an MCP server that any Claude Code-aware agent framework can call to escalate complex tasks to a local Claude Code instance — addressing a gap where sandbox-based agents can't access the real machine context Claude Code needs.