Garry Tan ac9291b146 evals: retry 1, not 2, on every paid path ... Measured on the llm-judge shard: --retry 2 amplified 25 tests into 46 executions (+84%), with retried runs at 138s vs a 10-12s baseline (429 backoff), and a permanently-failing test paying 3x. One retry still absorbs one-off flakes; chronic flakes become visible fix-work instead of silent wall-clock. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>		2026-08-15 07:53:39 -07:00
..
actionlint.yml	ci: right-size five single-core jobs to ubicloud-standard-2	2026-08-15 07:52:08 -07:00
ci-image.yml	ci: stop version bumps rebuilding the eval Docker image (cache key trio)	2026-08-15 07:49:54 -07:00
evals-periodic.yml	evals: retry 1, not 2, on every paid path	2026-08-15 07:53:39 -07:00
evals.yml	evals: retry 1, not 2, on every paid path	2026-08-15 07:53:39 -07:00
make-pdf-gate.yml	ci: fix workflow_dispatch concurrency collisions (head_ref || run_id)	2026-08-15 07:52:16 -07:00
pr-title-sync.yml	ci: right-size five single-core jobs to ubicloud-standard-2	2026-08-15 07:52:08 -07:00
skill-docs.yml	ci: right-size five single-core jobs to ubicloud-standard-2	2026-08-15 07:52:08 -07:00
version-gate.yml	ci: right-size five single-core jobs to ubicloud-standard-2	2026-08-15 07:52:08 -07:00
windows-free-tests.yml	ci(windows): cache bun installs; run the curated suite, not a hand list	2026-08-15 07:52:52 -07:00
windows-setup-e2e.yml	ci(windows): cache bun installs; run the curated suite, not a hand list	2026-08-15 07:52:52 -07:00