Skip to content

charlesjones-dev/ai-performance

v2.1.0MIT

Audits a codebase, or one directory, for performance bottlenecks and saves a timestamped report with an impact score for each finding, and measures where a repository's GitHub Actions time goes, with the file, line and seconds saved for each fix

performance-ci

Measure where a repository's GitHub Actions time goes and how to cut it. Pulls the last 150 completed runs with gh and splits every job into waiting for a machine, setup (checkout, installs, cache restore) and the job's own work. Lists the slowest jobs, then checks nine causes: runner queueing, slow lint, slow typechecking, unsharded tests or small machines, test files that reload the app, slow setup, jobs that don't run in parallel, full test runs on every PR, and no local checks before push. Each fix gets its file and line and the seconds it saves. Changes that cost money or run fewer tests go in a separate list for you to decide. Read-only until you pick the fixes.

Read SKILL.md at the source

Pinned to revision 94424cb2757a, so it is the text this page describes rather than whatever the author pushed since.

Files

Every link opens the file at its source, pinned to the revision this page describes.