langchain-ai/open-swe ? reverse-engineered prompt

Build me an open source coding agent platform that can take tasks from a web dashboard, GitHub, Slack, or Linear, work on them in its own isolated environment, and then open a pull request with the changes.

I want it to be able to understand a repo, make edits, run checks, and keep going on the same thread if I reply later. It should also be able to review pull requests, learn the style of a codebase from past feedback, and monitor CI so it can help with flaky failures or follow up on comments.

Please include a simple dashboard for starting and tracking work, plus the ability to run scheduled jobs. I also want a desktop mode for local projects if that fits. Keep it open source, configurable for different models and sandboxes, and safe by default with clear repo and permission boundaries. If you need to look up current docs online, go ahead.