[BUG] UltraReview timeout/crash consumed ~$37 in credits with no output returned

Resolved 💬 1 comment Opened Apr 29, 2026 by ethanortecho Closed May 31, 2026

Summary

Two consecutive UltraReview runs failed without producing any output, but both consumed credits. Total known charge: $37 (run 2). Run 1 charge unknown but non-zero.

This appears to be the same class of failure tracked in #49905 (timeout consumes quota) and #49882 (failed runs decrement free quota).

Environment

  • Claude Code: latest (UltraReview research preview)
  • Date: 2026-04-28
  • Account: eortecho@automized.com (Automized org)

Steps to reproduce

  1. Invoke UltraReview on a project from a Claude Code session.
  2. Observe behavior across two attempts.

Expected behavior

Either the run completes and returns findings, or it fails cleanly without consuming credits.

Actual behavior

Run 1: Crashed before doing any work. No output. Credit consumption unknown but non-zero.

Run 2: Run started successfully — could see it actively finding issues and verifying issues. Mid-way through verification, the run stopped. The parent Claude Code session reported a timeout. All in-progress findings were lost — no report, no partial output, nothing surfaced back to the user. This run consumed $37 in credits.

Impact

Roughly $37+ in credits charged with zero deliverable. Per the discussion in #49905, failed remote sessions should not decrement credits/quota.

Related

  • #49905 — ultrareview timeout shouldn't consume quota credits
  • #49882 — failed runs consuming free quota

Filing this for engineering visibility. Refund request submitted separately via support.claude.com.

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗