6 ms·
Here is my working flow, confirmed by more than 400 pr merged over the last 4 months. More than half of them were following my current strategy: 1. My agent wr
by alekstret 1mo ago
Here is my working flow, confirmed by more than 400 pr merged over the last 4 months. More than half of them were following my current strategy:
1. My agent writes code.
2. Then it creates tests and verifies that all of them actually work, not just pass. To do this, my agent writes the test, then it deletes the code it covers, reruns the test, confirms it goes red, and finally puts the code back.
3. I receive the ready-to-test code and environment setup.
4. I check that the business logic works as I expected it to be on a working product. Here we usually do several iterations of coding and bug fixing.
5. When the manual part is finished, the agent starts an external review using /code review skill. At that stage, it makes some additional fixes and corrections to the tests.
6. Finally, a branch is ready to be merged. We start CI/CD and wait until the run finishes successfully.
That's what I actually use because it generally works.
Note about only docs PRs:
I just ask the agent to make the changes, then it runs the / code review skill, and then we merge the branch into main without CI running.
- AlfeG 1mo agoI tend to do code review in separate session. So that context is not affecting judgments.
- ultrasandwich 1mo ago> To do this, my agent writes the test, then it deletes the code it covers, reruns the test, confirms it goes red, and finally puts the code back. Why? Are you aware of red/green/refactor?
- andned 1mo agono op but the tests shouldn't fail on a missing import. it proves nothing. I think what he means is that he makes sure that the test don't pass on testing something else pre-existing. i'm also struggling with vibe coded tests, sometimes they just test for random irrelevant stuff
- alekstret 1mo ago[dead]
- tarekabi 1mo ago[flagged]
- maleldil 1mo agoIsn't 1+2 just reverse TDD?