claude-code - 💡(How to fix) Fix [ultrareview] Two consecutive crashes on large diff — free credits consumed without findings

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…

/ultrareview crashed twice in a row on the same branch without producing any findings, but both attempts decremented my free credit counter (now at 0/3 with no actionable output).

Error Message

| 2 | r54lzf35u | https://claude.ai/code/session_015ydQ8XrY3wVFHggtpYMqoD | Same error | A multi-agent cloud review producing findings (or at minimum an actionable error message + a non-charged credit when the review fails internally). Same opaque crash twice, both billed against my free quota. The error message points to "session logs" that are not visible to me, so I cannot self-diagnose. 3. Improve the error surface: when the reviewer crashes, the CLI should either (a) not consume a credit, or (b) return something more actionable than "see session logs" (which the user cannot access).

Root Cause

/ultrareview crashed twice in a row on the same branch without producing any findings, but both attempts decremented my free credit counter (now at 0/3 with no actionable output).

RAW_BUFFERClick to expand / collapse

Summary

/ultrareview crashed twice in a row on the same branch without producing any findings, but both attempts decremented my free credit counter (now at 0/3 with no actionable output).

Crashed sessions

#Task IDSession URLResult
1r61isfvfhhttps://claude.ai/code/session_01PPbNuf27YUVAnZxjZWkSvhReview crashed before producing findings. See session logs for details.
2r54lzf35uhttps://claude.ai/code/session_015ydQ8XrY3wVFHggtpYMqoDSame error

Repro context

  • Branch: main
  • Scope reported by CLI: 4 files changed, 7145 insertions(+), 107 deletions(-)
  • Largest file in diff: public/style.css (~25k lines total, ~2500 lines added across the diff)
  • Project: vanilla HTML/CSS/JS Kanban app (no React, no build step)
  • Branch state: clean working tree before launch, both attempts launched within minutes of each other

What I expected

A multi-agent cloud review producing findings (or at minimum an actionable error message + a non-charged credit when the review fails internally).

What happened

Same opaque crash twice, both billed against my free quota. The error message points to "session logs" that are not visible to me, so I cannot self-diagnose.

Hypothesis

Either (a) the diff exceeds an internal context budget on the reviewer sub-agent and there is no graceful fallback, or (b) a transient infra issue that didn't recover on retry. In both cases, charging the user for a crashed run seems unintended.

Request

  1. Refund both free credits (`r61isfvfh` and `r54lzf35u`) — current counter shows 0/3 remaining but I got zero findings.
  2. Confirm the failure mode so I know whether to scope future runs to smaller diffs (single-file PRs) to avoid hitting the same wall.
  3. Improve the error surface: when the reviewer crashes, the CLI should either (a) not consume a credit, or (b) return something more actionable than "see session logs" (which the user cannot access).

Environment

  • Claude Code CLI on Windows 11 Pro (10.0.26200)
  • Shell: bash (Git for Windows)
  • Model: claude-opus-4-7 (1M context)
  • Working dir: a single-package vanilla JS project, no monorepo

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

claude-code - 💡(How to fix) Fix [ultrareview] Two consecutive crashes on large diff — free credits consumed without findings