What problem does it solve? Slow CI pipelines waste developer time, but most optimization attempts fail because teams guess at the bottleneck instead of measuring it. This Skill provides an ordered, eight-step method for finding where CI time actually goes — from step-level timing breakdowns through per-test duration analysis to verified A/B results — so you only fix what measurably matters. ## Core Features & Use Cases - Step-level timing breakdown: Read each job step's started_at/completed_at across several runs to identify the dominant step before profiling anything. - Bottleneck classification: Distinguish caching problems (setup/install dominance), out-of-process waits (subprocesses, sleeps, network calls), fixed per-test fixture costs, and CPU-bound versus waiting-bound suites. - Verified optimization discipline: A/B every change at suite level, revert what shows no measurable gain, and report local and CI numbers separately. - Use Case: Your test suite takes 10 minutes in CI. Following the method, you discover setup caching is fine, but a broad band of tests shares a slow fixture helper — fixing that one helper cuts the suite meaningfully, which you confirm with a real CI run. ## Quick Start Analyze my CI pipeline performance and tell me where the time actually goes and what is worth fixing.
