Software & DevOpsDone with Cursor Jan 1, 2025
Monitor tasks and open pull requests with long-running agents
A commenter runs a persistent script that watches tasks and uses Cursor to open pull requests, building a documentation summarizer and a data pipeline.
- What They Did
- In a comment, the team said they run a Claude Code server on a VPS using the Caliber AI setup repo. That setup generates MCP configs and agent skills for their codebase and Cursor, they sync the environment with one command, and then a persistent script watches tasks and has Cursor open pull requests; the examples they named were a documentation summarizer and a data pipeline.
- How It Went
- They said this workflow was used to build a few internal tools, specifically those two examples. The comment gives no metrics, timeline, or detail on review, so the results are described only at a high level.
- Worth Knowing
- The main implementation catch they mention is infrastructure: a VPS-hosted Claude Code server, generated MCP configs and skills, and a one-command environment sync.
Try it yourself with Cursor
In my repo [repo name], read the open tasks in [issue tracker or task file] and for the one I pick, make the code change on a new branch, run the tests, and open a pull request with a clear description. Start by listing the tasks you could handle and wait for my choice, and do not merge anything. Finish when the pull request is open and tests pass.