cowork-to-code-bridge — async bridge for local Claude Code execution from AutoGen agents #7868
abhinaykrupa
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Hey AutoGen community!
Building AutoGen agents that need to run code on a real machine (not a sandbox)? Thought this might be useful.
What it does
cowork-to-code-bridge is a local-first async bridge that lets your AutoGen agents escalate complex coding tasks to Claude Code on the user's actual Mac/Linux machine — then poll for results without blocking.
The pattern:
escalate_to_claude(prompt)via MCP → returnsoperation_idimmediatelyget_operation_status(operation_id)— non-blocking, idempotentresume_receiptstreamed back to the agentArchitecture
No cloud endpoints. No open ports. Auth by HMAC-signed token.
Production safety (recently shipped)
Why this matters for AutoGen
AutoGen's code execution defaults to Docker or local subprocesses — both of which are stateless. This bridge adds:
Full integration guide: https://github.com/abhinaykrupa/cowork-to-code-bridge/blob/main/docs/EXTERNAL_AGENT_INTEGRATION.md
Would love feedback from anyone building agents that need local machine access — especially around the stateful execution model. Does this pattern fit how you're using AutoGen?
→ Repo: https://github.com/abhinaykrupa/cowork-to-code-bridge
Beta Was this translation helpful? Give feedback.
All reactions