Repository navigation
Skip /go's fix-commit review and Final simplify for small, low-risk fixes - #271
Merged
Merged
Conversation
/go's single pass follows the rule, and the stage loop passes it to each stage implementer alongside the judgment-call rule. review-code --fix agents still get neither. ai/evals/smallest-change is an eval-only plugin for `claude plugin eval`. The with arm loads a SessionStart hook that prints rule.md, a copy of the rule that ai/tests/test-eval-rule-sync.sh keeps in sync. The without arm is the tool's no-plugin baseline, so the rule is the only difference. Three scaffolded Python cases grade with hidden tests, a production-line limit, a dependency check, and (in two cases) a check that the change uses the stdlib module or existing helper. Claude-Session: https://claude.ai/code/session_01AovXvYri48wFhqGV3epaMi
Judgment calls and Smallest complete change go to the single pass and each stage implementer only. The test writer and review-code --fix agents get neither. Claude-Session: https://claude.ai/code/session_01AovXvYri48wFhqGV3epaMi
…ixes review-tier.py decides over a commit range whether Review fix commits runs its review (skipped only for small docs or test changes with no risky path or uncertain change) and whether Final simplify runs (skipped for a few lines or docs only). The Report lists each skipped step. run-review.py copies the review child's token-usage rows into review-code's shared log, tagged with the go run IDs, so they survive removal of the worktree. Claude-Session: https://claude.ai/code/session_01AovXvYY2uZDJFssBEumPmd
haacked
force-pushed
the
haacked/go-scale-review
branch
from
October 29, 2026 16:09
59c9cf8 to
be56a78
Compare
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
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.
/goreviews its review fixes with a secondreview-coderun that costs about as much as reviewing the whole PR (a median of 1.5M tokens in review-code's log), even when the fixes are a few lines. Final simplify then runs over the same fixes. Each review's token rows also lived only under the worktree's.notes/, so they were deleted with the worktree and nobody could measure this.This skips those two late steps when the change is small and safe to skip. I kept the first review at full strength, because in review-code's log 4 of the 5 real blockers on diffs under 50 lines came from specialist reviewers, not correctness. Small code fixes keep their review too, since 7-line and 63-line deltas in the log each had a blocker. So
review-tier.pyskips the fix-commit review only for a small docs or tests change that touches no risky path (migration, API, wire format, auth, infra, dependency, SQL delete) and makes no uncertain change (rename, deleted file, binary, mode change). A file it doesn't recognize counts as code. Final simplify skips when the fixes change 20 lines or fewer, or only docs. The Report lists each skipped step, andrun-review.pycopies the review's token rows into review-code's shared log with thegorun IDs added.A separate review-code change will let
/gonarrow small code fix commits to correctness and testing instead.Test plan
python3 ai/skills/go/scripts/tests/test_review_tier.py, including a small migration, a rename across 12 files, deleted-only diffs,.txtand prompt files, and an added line starting with++python3 ai/skills/go/scripts/tests/test_run_review.py, including undecodable token rows and a symlinked shared log/goon a branch whose review fixes touch only tests, and check that the Report lists the skipped review and the unreviewed commitshttps://claude.ai/code/session_01AovXvYY2uZDJFssBEumPmd