Add /self:audit — Karpathy-style measure-first codebase audit - #25
Merged
Conversation
Systematic audit that measures everything before forming opinions: coverage per package, lint issues, security vulns, stale code, git health. Ranks hypotheses by impact/effort ratio with evidence citations. Suggests which devkit command to use for each hypothesis. Does NOT fix anything — only measures and analyzes. You decide what to test, one at a time.
- README tree: self-*.md (5) → (6) after adding self-audit - Remove fake Karpathy quote marks — reframe as paraphrase - Fix "What NOT to Do" contradiction: was implying fixing is in scope, now says "use the suggested devkit command" instead - Sync workflow "What NOT to Do" with command (4 bullets, same content) - Separate HAS_TS from HAS_JS detection (package.json alone ≠ TypeScript)
- Workflow detect step now includes {{input}} so user args aren't discarded
- Fix shell glob: [ -f */go.mod ] → find with -maxdepth 2 -print -quit
- Fix HAS_TESTS: now returns yes/no like all other variables
- Replace platform-specific stat with portable git log --format='%ai'
- README: "one-at-a-time" → "report only" (command doesn't fix anything)
The "quote" was fabricated. The actual inspiration is karpathy/autoresearch: one file, one metric, keep or discard, loop ~100 times overnight. We apply the same pattern to codebase quality instead of ML training. Now references the real repo with a link and describes the actual methodology.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
New
/devkit:self-auditcommand that implements the Karpathy research methodology: measure first, form hypotheses from evidence, rank by impact/effort, test one at a time.How it works
Key principle
Does NOT fix anything. Only measures and analyzes. The developer decides which hypothesis to test, one at a time. No bundling.
Use case
Run on any repo:
/devkit:self-audit→ get ranked improvement list → pick #1 →/devkit:self-improveor/devkit:self-test→ measure again.Test plan
/devkit:self-auditon devkit itself as a smoke test