Skip to content

charlesjones-dev/ai-improve

v1.0.0MIT

Audits a codebase across nine categories and writes self-contained implementation plans that another agent or a cheaper model can carry out, each with exact files, steps that end in a check, scope limits and STOP conditions. It also writes a plan from a description or from another audit's report, hands a plan to an executor in a separate git worktree and reviews the result, and keeps the plans up to date as the code changes. It never changes source code itself

improve-execute

Hand one implementation plan from plans/ to a cheaper executor model in a separate git worktree, then review its work the way a tech lead reviews a pull request: run every done criterion again, check that the diff stays in the plan's scope, read the diff and the new tests, and give a verdict of approve, revise (two rounds at most) or block. Never edits code itself, and never merges, pushes or commits to your branch.

Read SKILL.md at the source

Pinned to revision 6991774ab496, 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.