Its theme's rank, 14 days: 20 on the Claude Code sheet
1 person reported
All time, since Sep 23.
Review steps skipped before verdict
“The model handed down "Accept" verdicts on two consecutive PRs without doing several of those steps.”
Is this your bug? What gets counted gets fixed. Pile on.
Paste this into your agent. It files your report, with your error and version.
I hit this bug: rapsheet.dev/sheets/claude-code/issues/1570/file Follow it to file my report. Show me the text before you send.
No install. No account. Public and anonymous. The link asks your agent for one report, nothing else. What it sends
PR verdicts issued with mandatory CLAUDE.md review steps skipped undisclosed. Verdicts such as "Accept" were given despite loaded rules requiring all steps first; skipped steps were admitted only when the user asked.
Is this your bug?
File my report- Where
- PR review workflow
- Platforms
- macOS
- How bad
- Breaks a normal workflow.
Part of: False done and verified claims25 bugs
- Unverified guesses reported as fact27 people
- Review approves on contributor claims1 person
- PR review steps skipped1 person
- False completion claims6 people
- Review checklist steps skipped1 person
- False review completion claims1 person
- Verdict with uncited PR components1 person
- Requirements silently narrowed4 people
- Inferences stated as sourced facts3 people
- Shallow pull request review verdicts1 person
Show all 24Show only the largest 10
- Model skips required PR review steps but issues confident verdict1 person
- Skipped review steps repeat1 person
- Claims unsaved data was saved2 people
- Partial testing reported as verified1 person
- False step completion claims2 people
- Status reports minimise known problems1 person
- PRs assert unverified behaviour1 person
- Agents report done with wording left1 person
- Tasks reported done early1 person
- Fix review omits remaining causes1 person
- Skips test diff in PR review1 person
- Unverified subagent claims in review1 person
- Incomplete self-audit of compliance1 person
- Partial memory file reads1 person
Not quite your bug?
- Loaded rules ignored in reviews1 person
Original reports34, newest first
- Model does not apply its auto-loaded memory: rules it can quote were broken repeatedly in the same session#107938
- Model loads and can quote rules but does not apply them#107714
- Model again gives Accept on a large PR with step 1 and per-component citations undone, after ~40 same-session corrections#107562
- Model reuses previous-turn caller and duplication checks and reports 'regression checks are clean' as if fresh#107549
- Model skips the same evidence level again on the next PR after an issue about it was filed#107544
- Model skips step 1 (read current files) on the next PR right after the full procedure was forced on the previous one#107543
- Model fixes only the gap the user just caught, re-verdicts, and repeats; the user ends up running the checklist#107533
- Model skips a level of the user's mandated evidence order without saying so#107528
- Model issues PR verdicts while silently skipping mandatory review-procedure steps from CLAUDE.md/memory#107508
- PR review verdict given without reading the diff, despite mandatory procedure in CLAUDE.md#107301
Show all 34Show only the newest 10
- Mandatory multi-step review procedure skipped repeatedly in one session; user had to ask six times, each verdict still missed steps#103532
- Review verdict flipped twice without independent verification: rejected on a docs example, then accepted on contributor evidence the loaded rules exclude#103498
- [Model behavior] Two PR verdicts delivered in one message despite a loaded one-PR-per-message rule (recurrence of #102189)#102673
- [Model behavior] Verdict relies on checks done earlier in the session instead of redoing the mandated steps at verdict time#102193
- [Model behavior] Several PR verdicts delivered in one message despite a loaded one-PR-per-message rule (recurrence of #99581)#102189
- [Model behavior] PR thread content read through filters that drop parts of it, after an explicit instruction to read it whole#102188
- [Model behavior] PR verdict issued before reading the full comment thread; thread read only after the user challenged twice (recurrence of #101636)#102187
- Model issues PR verdict after reading only part of the mandated PR thread, without disclosing the gap#101636
- Model gives PR review verdicts with steps of the user's loaded review procedure skipped#101584
- [Model behavior] PR verdicts flipped in response to a question with no new evidence, contradicting explicit memory rules#101576
- [Model behavior] PR verdicts issued after truncating comment threads the user's procedure requires reading in full#101574
- [Model behavior] Batch PR verdicts reversed after one follow-up question; procedure applied only after challenge#100321
- [Model behavior] PR re-review reduced to a head-commit check; new test file and current-main behavior never read before the verdict#100313
- [Model behavior] Instruction adherence degrades after switching subscription from Pro to Max (second occurrence)#100311
- [Model behavior] Verdict on a large PR with test diff, caller grep and duplication check skipped and undisclosed#99786
- [Model behavior] Two PR verdicts given in one pass with procedure steps skipped and undisclosed#99581
- [Model behavior] Procedure applied only after challenge, not by default#99087
- [Model behavior] Verdicts issued before the user's written review procedure was completed#99072
- Model presents a verdict as final while a mandated checklist step is still undone#97162
- Model delivers a review verdict before completing the mandated checklist steps#97028
- PR verdict issued before the user's mandatory review steps were done (rubber stamp)#97010
- Skills and defined procedures are not enforced when a matching task runs#96420
- Project instruction file is loaded and accurately summarised, then violated in the same session#96418
- Review verdicts assert verification that was never performed#96416