/ultrareview: two consecutive free runs failed (empty findings, then crash) — requesting credit restoration

Invoked /ultrareview 8 twice against PR oasys-works/oecs#8 (a
v0.3.0 release PR: 60 files changed, +4,379 / −2,163, substantial
breaking changes in public APIs). Both runs consumed a free-tier
use and produced no usable output.

Run 1

• Session: https://claude.ai/code/session_01L5AkJAXXH4etuyY8eU23bD
• Target: 2khan/oecs#8. Local origin at the time was

git@github.com:2khan/oecs.git, a GitHub redirect alias for the
renamed/transferred canonical repo oasys-works/oecs.

• Outcome: completed normally, returned empty findings [].
• Expected: a release-sized diff with breaking API changes should

produce some findings; zero is almost certainly a silent failure.

Run 2

• Session: https://claude.ai/code/session_01XbWWkXSzvno7Jdrw3cszpz
• Target: oasys-works/oecs#8. Between runs I updated the local

remote to the canonical URL (git remote set-url origin
git@github.com:oasys-works/oecs.git) so the redirect was no
longer in play.

• Outcome: returned `{"error": "Review crashed unexpectedly. See

session logs for details."}`

So the two failures have distinct characteristics: run 1 may or may
not have been related to the redirected remote, but run 2 crashed
against the canonical URL with no ambiguity.

Possible product gap (independent of the crash)

If the empty-findings output in run 1 was caused by the redirected
remote, it'd be worth having /ultrareview either follow GitHub's
repo redirect (as gh and git do transparently) or surface a
clear error so the user can correct before burning a quota use.

Ask

Please restore both /ultrareview quota credits. Happy to provide
additional detail from either session.