GitHub Workflow Reliability and Flaky Test Audit
Find unreliable checks that slow everyone down.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Audit CI reliability. Data from the last 100 workflow runs: job name, duration, result, re-run count and failure messages: [RUN_DATA]. Present results per job in a table with runs, failures, re-runs that passed, median and 90th percentile duration, and a flakiness estimate: failures that passed on re-run divided by all failures. Worked example: 18 failures of which 11 passed on re-run is 11 / 18 = 61 percent, which suggests flakiness. State the sample size. Identify the three least reliable jobs and likely causes. Quantify the cost: 18 failed runs of 12 minutes plus re-runs of 12 minutes is 432 minutes, about 7 hours of waiting, and give a cost estimate in minutes per week. Propose actions: quarantine flaky tests with an owner and a fix date (do not hide them by deleting), add retries only with logging, and speed up with caching. Return a table, a dashboard of three numbers, and a weekly 15-minute review of the worst job.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.