I Tested Every Cursor Trick—So You Don’t Have To!
summary
Hands-on report of using Cursor as primary coder: it fixes an agenda submit bug, converts a formatTime function into a tested util by installing testing libraries and running TDD, and generates PR descriptions and commit messages, but it also sometimes breaks the build and leaves failing tests. Key features used are agent/ask/manual modes, add docs and at web scraping, and add git diffs for pinpointing regressions. Auto-run and auto-fix lint speed iteration but require project and user rules to avoid unwanted installs or edits.
AI-written summary · about
chapters (15)
- 0:00Introduction and setup
- 1:48Cursor modes and basics
- 3:58Agent mode demo
- 5:47Writing tests with AI
- 7:11Adding context and web search
- 9:19Docs and internal docs
- 10:04Using add diff for debugging
- 11:27Handling test failures
- 14:17Working with images
- 15:47Setting boundaries with rules
- 17:55User and project level rules
- 19:23Generating rules automatically
- 20:27Key tips for effective use
- 21:12Additional cursor tricks
- 22:20Final thoughts and Q&A
related talks
37mAutomated Performance Testing With WebDriver
32mLiving on Crotches or UI Automation Antipatterns
40mBenjamin Bischoff: Old Tools, New Tricks: Unleashing the Power of Time-Tested Testing Tools
47mModern Performance Workflows
37mFront-end Testing for Skeptics
53mWait, I have to test the front end too?
21mWorking Well, The Future of JavaScript Testing
39mATDD with CodeceptJS: Effective Testing Without Pain