Run asynchronous coding-agent workflows with Open SWE
Deploy an internal coding agent that accepts tasks from GitHub, Slack, or Linear, runs them in isolated sandboxes, and opens pull requests automatically.
npx skills add agentskillexchange/skills --skill run-asynchronous-coding-agent-workflows-with-open-swe
https://github.com/langchain-ai/open-swe, run uv venv, source .venv/bin/activate, and uv sync --all-extras, then follow the installation guide to configure the GitHub App, LangSmith OAuth, sandbox snapshot, and webhook triggers.Use Open SWE when a team wants an asynchronous coding agent for repository tasks that should run outside a developer’s local IDE. The operator configures the GitHub App, LangSmith, sandbox provider, and trigger integrations, then routes issues or comments to Open SWE so it can clone the repo, work in an isolated environment, coordinate subagents, and open a draft pull request. Invoke this for supervised internal coding-agent queues, not for one-off local edits or generic LangChain examples.
What this skill actually does
Inputs and prerequisites: Python, uv, GitHub App, LangSmith, sandbox provider, GitHub/Slack/Linear trigger.
Setup notes: Clone `https://github.com/langchain-ai/open-swe`, run `uv venv`, `source .venv/bin/activate`, and `uv sync –all-extras`, then follow the installation guide to configure the GitHub App, LangSmith OAuth, sandbox snapshot, and webhook triggers.
Source and verification boundary: use https://openswe.vercel.app as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Custom Agents workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.