DottaandClaude Opus 5.5 26e77362c5 fix(evals): grade steering.steer-active honestly on runtimes that cannot steer ...

The OpenCode harness reports steering: false, so the queue for its running
task says steeringDisposition "unsupported" and the page offers Interrupt
instead of Steer. The fixture waited only for "available" and timed out.

Accept either definitive capability. When steering is available, the
same-turn ACK branch is graded as before. When it is unsupported, grade the
honest fallback the scenario names: no usable Steer control, a visible
continuation control, a refused steer attempt (409 steering_unsupported) that
records no steering and keeps the exact message queued, and one normal
continuation that consumes it and renders the blue page. Drop the reply
keyword checks and the /blue/ match on the delivery document; the rendered
page proves the colour.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-11-11 12:23:18 -06:00

..

aggregator-accounts

feat(connections): sync and group accounts managed by aggregators (#15254)

2026-10-05 14:57:04 -05:00

ai-connections-app

feat(connectors): show and select personal AI account defaults (#19845)

2026-11-07 10:26:08 -06:00

ai-connections-review

fix(build): typecheck package, Playwright, and e2e test files in CI (#18839)

2026-11-06 07:33:59 -08:00

canary-onboarding

test(release): verify paused onboarding setup records (#18102)

2026-10-28 16:03:08 -07:00

e2e

perf(task-chat): open tasks from a preloaded cache and cold tasks in one request (#20592)

2026-11-10 14:26:27 -06:00

fixtures

test(e2e): reuse core journeys across deployment targets (#18222)

2026-10-29 18:58:26 -05:00

helpers

feat(cases): add persistent workflows with same-task agent handoffs (#16959)

2026-10-20 12:46:14 -05:00

hiring-ai-connections

fix(runner): enable API tools by default (#14186)

2026-09-26 21:16:33 -05:00

html-preview

feat(ui): render HTML artifacts in a secure sandbox (#15447)

2026-10-07 09:01:05 -05:00

lifecycle-baseline

fix(build): typecheck test projects in 4 GiB (#18738)

2026-11-03 11:09:12 -08:00

perf

perf(task-chat): open tasks from a preloaded cache and cold tasks in one request (#20592)

2026-11-10 14:26:27 -06:00

release-smoke

test(release): verify paused onboarding setup records (#18102)

2026-10-28 16:03:08 -07:00

restart-chaos

fix(heartbeat): retry every pre-provider lease acquisition failure and rest it on Retry (#20857)

2026-11-11 12:05:54 -06:00

runner-acceptance

feat(runner): consolidate Cursor production integration (#15075)

2026-10-06 20:48:15 -05:00

runner-e2e

fix(evals): grade steering.steer-active honestly on runtimes that cannot steer

2026-11-11 12:23:18 -06:00

runner-recovery

test(runner): add everyday workflow evaluation harness (#13474)

2026-09-15 11:04:16 -05:00

speko-native

fix(build): typecheck test projects in 4 GiB (#18738)

2026-11-03 11:09:12 -08:00

speko-voice-proof

test(voice): qualify Speko task conversations and Storybook surfaces (#15729)

2026-10-13 16:54:41 -05:00

storybook-visual

fix(links): retain PR status across page reloads (#20654)

2026-11-10 17:43:06 -06:00

acpx-qualification-models.ts

feat(codex): default Codex to gpt-6.1-sol, half the gpt-5.6-sol price (#19623)

2026-11-06 15:35:34 -06:00

tsconfig.playwright.json

fix(build): typecheck package, Playwright, and e2e test files in CI (#18839)

2026-11-06 07:33:59 -08:00