From 3f02efdef98fad6ef074ebd978d0e8509d7fd178 Mon Sep 17 00:00:00 2001 From: Danny Avila Date: Thu, 30 Jul 2026 13:44:36 -0400 Subject: [PATCH] =?UTF-8?q?=E2=9A=A1=20feat:=20Interrupt=20&=20Steer=20(In?= =?UTF-8?q?itial=20UI)=20(#14528)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit * πŸ›‘ feat: Preemptive Steer β€” server half (Interrupt & Steer, PR 2 of 3) Lets the steer route ask the generating replica to seal its live model stream at the next provider-safe boundary instead of waiting for a tool step. The run is never aborted, job status never changes, the partial answer is kept, and generation resumes in the same assistant message after the injected steer. Consumes the SDK seam in @librechat/agents (danny-avila/agents#335, #346). Transport: IEventTransport gains a fenced emitPreempt/onPreempt pair beside abort. RedisEventTransport fans PREEMPT out on the SAME events channel and subscription (no new connection, key, or subscribe call); onPreempt returns a registration-scoped unsubscribe with the same replacement-safe state-identity guard onAbort uses. InMemory implements neither β€” single-process preempt lives entirely in the runtime set. Runtime state: RuntimeJobState carries the per-generation request set, createdAt-fenced and capped at STEER_QUEUE_MAX_DEPTH, plus a bounded `cleared` tombstone so a late cross-replica arm cannot resurrect a request whose steer already drained. registerPreemptSubscription mirrors the abort registration's double fence (runtime identity + generation createdAt); releaseAbortSubscription retires BOTH listeners and the armed set, so every terminal path drops preempt state for free. Public surface: requestPreempt (arm + fenced publish, never a rejection surface, never touches job status), isPreemptRequested (O(1) level-triggered poll), noteSteersRemoved (drain/cancel bookkeeping + fenced clear), clearPreemptRequests (empty-boundary disarm). One drain body, two boundaries: createSteerDrainHook (PostToolBatch) and createSteerPreemptBoundaryHook (PreemptBoundary) share drainAndBuildInjections, so the two injection sites cannot drift β€” the SDK's provider-safety argument rests on identical HumanMessage shapes. The shared body builds injections incrementally under a swallow-all catch (a mid-loop throw still injects what was applied β€” those parts are already persisted), clears preempt requests in finally, and disarms the generation when a boundary drains nothing. Request path: POST /chat/steer accepts preempt: true. The guard ladder is unchanged in order and in every status code. A preempt request is NEVER a rejection reason β€” without the capability the steer still enqueues and the 202 echoes preempt: false. Armed strictly after a successful enqueue; cancel disarms. The capability is read from the OWNING replica's recorded `preemptCapable` rather than the route replica's own SDK probe, so a rolling deploy cannot label a steer "interrupting" that the older owner will only inject at a tool step. Durable label: SteerQueueItem.preempt β†’ TPendingSteer.preempt, so a parked/claimed/replayed chip keeps its wording. Run wiring: createRun registers the PreemptBoundary hook and threads RunConfig.preemption, both gated on isSteerPreemptSupported() β€” a separate probe from isSteeringSupported(), so the client affordance can never arm against an SDK that only injects at tool boundaries. buildSteerWiring builds both hooks from one shared closures object, so preemption survives HITL pause/resume for free. Honest finalization: an empty preempt boundary persists and emits with unfinished: true β€” the same contract an abort gets β€” re-marked explicitly because BaseClient has already saved the row as unfinished: false by that point. Not changed: no new job status, store method, Lua, SSE event type, endpoint, or authorization surface. abortJob, completeJob, transitionStatus, closeAndDrainSteers, getResumeState, emitChunk, applySteerPart and the whole abort path are untouched. Tests: 120 packages/api steering specs (preempt lifecycle, tombstone, fences, caps, terminal release, both-boundary drain parity, level-triggered poll, request/cancel arming, owner-capability degradation) plus 5 in api for buildSteerWiring gating, and 2 Redis-gated cross-replica transport specs. * πŸ”’ fix: Codex round 2 β€” evict tombstones, scope the empty-boundary disarm, honest resumes All four server findings were fresh consequences of the round-1 fixes, which is the review doing exactly what it should. - Tombstone cap refused new entries instead of evicting. Every drained or cancelled steer is tombstoned, not just preempting ones, so a generation that processed 20 steers exhausted the set and the late-arm race resurfaced silently. Now evicts oldest-first (Set iteration is insertion-ordered), with the budget named PREEMPT_TOMBSTONE_MAX rather than an inline expression. - The empty-boundary disarm I added in round 1 wiped the generation's ENTIRE armed set. A second steer can enqueue and arm between the atomic drain returning empty and the disarm running β€” that arm is backed by a live, uninjected queue item and must survive. The drain now snapshots the armed ids BEFORE draining (getArmedPreemptIds) and clearPreemptRequests takes an explicit id list instead of clearing everything. - HITL resume finalized with a hardcoded unfinished: false. The boundary hook is re-registered on resume via buildSteerWiring, so a resumed segment can end on an empty preempt boundary exactly like a fresh one; finalizeResumedTurn now reads getPreemptStats() and the halt reason, matching the normal request path. - Ownership moves on resume, so the job's recorded preemptCapable must describe the replica that will actually generate. Refreshed before resumeCompletion; a job created on a capable replica that resumes on an older one during a rolling deploy no longer acknowledges steers as interrupting. Tests: +3 (scoped disarm sparing a post-snapshot arm, oldest-first tombstone eviction, id-list disarm). 122 packages/api steering specs green. * 🚨 fix: Codex round 3 β€” deserialize preemptCapable from Redis (feature was dead under Redis) The P1 here is the most consequential defect in the whole feature, and it was introduced by round 1's own capability fix. - `RedisJobStore.serializeJob` writes booleans generically, so `preemptCapable` reached Redis β€” but `deserializeJob` is an EXPLICIT field map and had no line for it. Every `getJob()` therefore dropped the flag, `job.metadata.preemptCapable` was always undefined, and `handleSteerRequest` computed `preemptArmed: false` unconditionally. Interrupt & steer would have silently degraded to ordinary tool-boundary steering in EVERY Redis deployment β€” i.e. the feature shipping as a no-op in production while passing every in-memory test. Now deserialized, with a round-trip assertion in the metadata spec that fails (`Received: undefined`) against the unfixed store. - The resume capability refresh moved from just-before `resumeCompletion` to immediately after `approvals.resolve` claims the run. That call already flips the job back to `running`, so the steer route accepts requests from that instant; leaving the refresh 135 lines later (across the whole client reconstruction) left a real window where a steer read the PREVIOUS owner's capability. Not the fully atomic transition Codex suggested β€” that reaches into the approvals Lua β€” but it shrinks the window from seconds to one await, which is proportionate for a label-accuracy issue. Refuted: "avoid triggering preemption inside subagents". The premise β€” that the run-wide poll can seal a subagent stream β€” does not hold against the shipped SDK. Child graphs are constructed with `subagentScope: true` (SubagentExecutor) and `preemption` is NOT propagated into child inputs, while `canClaimPreemptSeal()` requires `!subagentScope && preemption != null`. Both conditions fail independently, so a subagent can never claim a seal and the boundary cannot fire with `agentId` set. The `input.agentId != null` guard in the hook is defensive depth, not the thing standing between us and the described failure. 140 packages/api specs green. * πŸ” fix: Codex round 4 β€” re-arm durable interrupt steers when resume moves owners - An arm lives only in the owning replica's runtime plus a transient pub/sub message, while the steer's `preempt` flag is durable on the queue item. A HITL resume landing on a different replica therefore started with an empty armed set and a poll stuck false, so an interrupt the user had already been ACKed for silently waited for an ordinary tool boundary. New `GenerationJobManager.rearmQueuedPreempts` rebuilds the armed set by peeking the durable queue (fenced on the generation) and re-arming every item flagged `preempt`; resume calls it right after claiming. Safe by construction: every item peeked is still queued, so no drained steer can be resurrected. - Capability-refresh failure now logs at error rather than warn, but deliberately does NOT fail the resume β€” see the reply on that thread. Tests: +2 (rebuild from queue arms only the flagged item and reports the count; a stale generation arms nothing). 124 packages/api steering specs green. * πŸ“‘ fix: Codex round 5 β€” acknowledge only what was actually armed - A cross-replica arm was fire-and-forget: `emitPreempt` logged its own publish failure and `requestPreempt` returned void, so the route answered `preempt: true` even when the owner never armed a poll. The steer still injected at the next tool boundary, but the chip claimed an interrupt that could not happen β€” and unlike HITL resume, an ordinary running generation had no durable reconciliation to recover it. `emitPreempt` now resolves to the subscriber count and rejects on failure; `requestPreempt` is async and returns whether the arm truly landed (owned locally, or delivered to at least one subscriber). The 202 reports THAT rather than what was asked for, so the chip relabels to ordinary steering exactly as it does for a capability-degraded deployment. Errors are swallowed into `false` β€” an unarmed interrupt is a downgrade, never a failed steer. - The owner capability is re-read immediately before enqueue rather than reused from the top of the guard ladder. `checkAgentAccess` and file resolution are awaits, so a request can span an entire HITL pause/resume that moves ownership to a replica with different capability and rewrites that very flag. Only paid for by requests that actually asked to interrupt. Tests: +3 (not-armed when the publish reaches nobody; armed when this replica owns the generation; a throwing publish downgrades instead of propagating). 127 packages/api steering specs green. * 🎯 fix: Codex round 6 β€” real ownership, confirmed disarms, and a CI regression of my own Three review findings plus three CI failures the round-5 commit caused. Review: - Ownership came from `runtimeState`, which a cross-replica `getJob` populates with a FACADE runtime on any replica that merely read the job. Matching `createdAt` therefore proved only "we looked at this job", so a non-owner could arm nothing and report success. Ownership now comes from `ownedJobs`, the actual owner map. - `armPreemptIds` returns how many ids it accepted, and a local arm is only reported as armed when one was. A tombstoned id (its steer drained at an ordinary boundary mid-request) no longer answers `preempt: true` for an interrupt that cannot happen. - The cancel disarm is awaited. A dropped clear is worse than a dropped arm: the owner keeps a level-triggered request for a steer that no longer exists, seals its next chunk and truncates an unrelated answer. The boundary drain's own call stays non-blocking β€” there the owner is local, so the disarm is already effective and awaiting the informational publish would only delay injection. - Subscriber count is NOT read as proof of owner receipt: the count includes this replica's own facade subscription. A successful publish reports armed, a rejected one does not. Documented rather than papered over β€” see the acknowledgement-semantics note on the PR. CI regressions from round 5, all mine: - `registerPreemptSubscription` was AWAITED at both runtime-init sites, so job creation blocked on a second Redis channel subscription and hung when that subscribe was slow. Abort is awaited because a missed abort strands a run; a missed preempt only degrades that steer to the next tool boundary, so it now registers without gating createJob. - Two api specs mocked `@librechat/api` without the newly imported `isSteerPreemptSupported`, so the call threw before createJob; and one exact-match assertion needed the new `preemptCapable` metadata field. - My own Redis integration spec asserted arm-before-clear ordering, which two publishes carry no guarantee of β€” the receiving tombstone exists precisely because of that. Now asserts delivery and payload fidelity, order-independent. 158 packages/api specs, 27 api specs green. * 🧭 fix: Codex round 7 β€” settle the acknowledgement semantics (Option A) Round 7's second finding is the incoherence I flagged on the PR: the route persisted `preempt: true` on the durable queue item while returning `preempt: false` when delivery could not be confirmed. Those two then disagreed, and `rearmQueuedPreempts` trusts the DURABLE one β€” so a resumed owner would honour an interrupt the client had explicitly been told degraded to ordinary steering. Rather than patch the disagreement, this settles the meaning: `preempt` in the 202 means "queued as an interrupt request", NOT "a seal is guaranteed". It mirrors `SteerQueueItem.preempt` exactly, so the response, the durable record, and the resume-time re-arm can never disagree. The gates that ARE knowable stay β€” the owner's recorded capability and a successful enqueue. Everything past that degrades to the documented fallback of injecting at the next tool boundary. A route cannot synchronously know whether another replica will seal: proving it needs a correlated request/response over pub-sub, and even that only proves the owner heard, not that it is still streaming when the arm lands. Four rounds of tightening this boolean each surfaced a narrower case; the sequence does not converge, so the invariant is now "the flag describes the durable decision" and an unconfirmed arm logs a warning instead of rewriting the answer. Also from this round: a failed disarm publish is retried once and its outcome reported. `handleSteerCancel` keeps `removed: true` β€” the steer really did leave the queue, and saying otherwise would make the client re-show a chip for a steer that can never arrive β€” and adds `disarmed: false` so the residual risk is visible rather than swallowed. Damage stays bounded regardless: the empty-boundary self-clear disarms the generation after a single seal. Tests: +1 pinning the response/durable-flag invariant. 159 packages/api specs green. * 🧹 fix: Codex round 8 β€” remove the unverifiable disarm signal Round 8 found the same over-promise on the disarm side that round 7 corrected on the arm side, so this applies the same answer rather than patching around it. The `disarmed: false` field added in round 7 was both unreliable and unused: a resolved publish is not proof the owner heard it (the delivery count includes this replica's own facade subscription), and it was never threaded into `CancelSteerResponse` or read by any client. A signal that claims a certainty the transport cannot provide is worse than no signal β€” it invites callers to trust it. Removed from the response. The retry stays, because it genuinely reduces the failure rate, and `noteSteersRemoved` still returns whether the publish succeeded FOR LOGGING, now documented explicitly as "published without error", not "the owner disarmed". Disarm is best effort with a bounded, self-healing failure: if the clear is lost the owner seals once, the empty-boundary self-clear disarms the generation, and the turn is persisted `unfinished: true` rather than silently truncated. Tightening that further needs a correlated request/response over pub-sub with a timeout β€” noted on the PR as the deliberate boundary of this design rather than an oversight. 130 packages/api steering specs green. * 🧽 fix: Codex round 9 β€” spend snapshot arms on nonempty drains too The round-6 scoping fix only cleared the pre-drain snapshot when the drain came back EMPTY. On a nonempty drain the `finally` cleared just the drained ids, so a stale arm β€” typically a cancel whose cross-replica clear was lost β€” survived the boundary. It would then immediately seal the continuation meant to answer the steer that had just been injected, and land on an empty boundary as `preempt_incomplete`: the interrupt appears to work, and the answer to it is truncated. A boundary that runs has spent its seal, so everything armed at snapshot time is spent whether or not it came back from the drain. The `finally` now clears the union of the snapshot and the drained ids. Arms that land AFTER the snapshot are still spared β€” their queue items are live and uninjected, which is the property round 6 added. Also fixes an api-workspace CI failure of mine: `resume.spec.js` stubs `GenerationJobManager` wholesale, and the round-3/4 resume work added two calls (`updateMetadata`, `rearmQueuedPreempts`) the stub did not define, so 34 specs threw. Stub extended. Tests: +2 (a nonempty drain clears a stale snapshot arm; a nonempty drain spares an arm that landed mid-drain). Counterfactually verified β€” the stale-arm spec fails against the unfixed drain. 132 packages/api specs, 60 resume specs green. * fix: never let a failed preempt subscription reject into the void registerPreemptSubscription is called detached at both sites, so a rejected Redis SUBSCRIBE became an unhandled rejection β€” process-fatal under Node's default --unhandled-rejections=throw. The comment already promised this path merely degrades steering; it now does. Swallowed and logged inside the registration rather than at each call site, so a future third caller cannot reintroduce the trap. Losing the channel costs this generation's cross-replica preempts, not the server: same-replica arming is runtime state and still works, and remote arms fall back to the next tool boundary. Verified counterfactually β€” the new spec surfaces SUBSCRIBE failed as an unhandled rejection against the unfixed registration. * docs: state the real blast radius of a failed preempt subscription LibreChat's own entrypoints install a global unhandledRejection handler that logs and keeps serving, so the escaping rejection this guards was never fatal to this server β€” only to another consumer of @librechat/api that installs no handler. The fix stands either way; the comment just should not overstate what it prevents. * test: cover the cross-replica preempt hop with two manager instances Every other preempt test runs against a single manager, so the hop that actually carries an interrupt in production had no coverage: the steer POST lands on whichever replica the balancer picks, which is usually not the one generating. Non-owner publishes, owner arms, owner's level-triggered poll flips β€” none of that was exercised end to end. Two GenerationJobManagerClass instances are a faithful replica pair here. runtimeState and ownedJobs are private instance fields, there is no module-level mutable state between them, and createStreamServices duplicates a dedicated subscriber connection per call, so separate OS processes would exercise the same objects over the same Redis. Both assertions verified counterfactually against real Redis: - Deleting the preemptCapable deserialization in RedisJobStore fails this with 'Expected: true, Received: undefined' β€” the exact P1 that shipped past every in-memory test and would have made the feature a silent no-op on every Redis deployment. - Dropping the non-owner arm publish fails it with 'Received: false'. * test: remove the fixed sleeps and vacuity from the cross-replica preempt test Codex round 11, both findings, both on the test I added last commit. P2 β€” the 300ms waits were load-bearing. Redis pub/sub never replays and the owner's SUBSCRIBE is detached, so on a slow CI worker the publish could land before anyone was listening and the test would fail against correct code. Now it republishes until the owner's state converges, which is safe because arms and clears are idempotent set writes keyed by steerId. Side effect: the tests got ~10x faster (85ms/57ms vs 929ms/606ms) since they finish on delivery rather than on a timer. P3 β€” afterEach destroyed only the transports, leaving each manager alive in its own cleanup-interval closure, still working against a dead transport. Now tracks the managers and awaits destroy(), which disposes the job store and its timer too. Matches how the rest of this file cleans up. Fixing the sleeps exposed a third problem codex did not flag: the stale-arm test could pass vacuously, because an undelivered arm and a fenced one look identical. It now brackets the stale publish between two control arms β€” the first proves the owner is listening before the stale one is sent, the second proves it has had its chance to arrive. Verified counterfactually against real Redis, and stable over 5 runs: - dropping the preemptCapable deserialization fails with 'Received: undefined' - dropping the non-owner arm publish times out both tests - removing the generation fence fails the stale test with ["control-before", "steer-stale", "control-after"] β€” which also confirms the bracketing orders as intended rather than by luck * fix: gate interrupt on the OWNER's capability alone, not the route's Codex round 12. The comment above this gate already said 'the OWNER's recorded capability, not this replica's probe' β€” and then the code ANDed in isSteerPreemptSupported(), which is exactly this replica's probe. The contradiction dates to the original commit; round 6 made the gate owner-scoped and wrote that comment without removing the local conjunct. The route never seals. It enqueues and publishes an arm, neither of which touches the SDK, so during a rolling deploy a steer landing on an un-upgraded replica silently lost its interrupt even though the owner could seal. When the route IS the owner the probe is redundant anyway: the flag it would consult is the one this process wrote at createJob. The real degradation path is unchanged and still tested β€” an owner that recorded no capability relabels to an ordinary steer. The test that pinned the local probe asserted an impossible same-replica state (capable metadata plus an incapable local SDK, when the metadata is written from that probe); it now pins the mixed-SDK direction instead, and fails with 'Expected: true, Received: false' if the probe is put back. * fix: reconcile arms at handover, and stop holding the 202 on a publish Codex round 13, two of three findings. P2 β€” rearmQueuedPreempts only ever ADDED. A replica that merely read the job still installs a facade runtime and subscribes, so it can accept an arm and then miss the best-effort clear that follows the drain. HITL resume promotes that facade to owner, the union keeps the orphan, and the first resumed stream seals on a steer no longer in the queue, drains nothing, and truncates the resumed answer as preempt_incomplete. acquireResumedJobOwnership only sets ownedJobs, so nothing else was clearing it. The durable queue is the sole authority at a handover: arms it does not back are now disarmed and tombstoned, so an in-flight publish cannot revive them either. Worth recording that my own independent review raised this and my verifier refuted it. Codex found it separately; two reviewers converging should have outweighed one refutation. P2 β€” the route awaited the arm publish before answering. The 202 reports capability, not delivery, so the await could not change the response; it only exposed the caller to Redis latency after the queue item was already durable. A client that times out and retries mints a second steer while the first stays queued, injecting the same instruction twice, whereas a lost publish merely takes the tool-boundary fallback. Detached, with both outcomes logged. All three tests verified counterfactually: union-only rearm fails the two new handover specs, and re-awaiting the publish hangs the stalled-publish spec until jest kills it. * fix: snapshot arms before reading the queue at handover Codex round 14 β€” a regression from my own round-13 fix, and a worse failure than the one it corrected. Round 13 read the durable queue first, then tombstoned any armed id the snapshot did not back. But approvals.resolve reopens steering before reconciliation runs, so another replica can commit a preempt steer and publish its arm while the peek is in flight. That arm is then present locally but absent from a snapshot taken before the steer existed, so a LIVE interrupt the route already acknowledged got dropped β€” and tombstoned, which blocks the re-arm, making it unrecoverable rather than merely late. Fixed by inverting the two reads rather than by locking or paying a second round trip. A steer is durably enqueued BEFORE its arm is published, so any id in an arms-first snapshot was already queued when it was armed, and the later peek must observe it unless it has since drained β€” which is exactly the orphan this reconciliation exists to drop. Arms landing after the snapshot are simply not candidates. Also re-checks runtime identity across the await, since the generation can be replaced while the queue read is in flight. New spec injects a steer + arm during the peek and verifies it survives; against the round-13 ordering it fails with Received array: []. * fix: bound the cancel disarm wait and fence enqueue to its generation Codex round 15. P2 β€” the cancel awaited its disarm publish unbounded. ioredis queues commands during an outage rather than rejecting, so that await could hang for the length of the outage with the steer ALREADY durably cancelled; a client that gives up then treats the cancel as failed and restores a chip for a steer that can never produce an applied event. Every successful cancel publishes, so ordinary steers were exposed too, not only preemptive ones. Now bounded at 1s, with the publish continuing behind it β€” its retry and logging are unchanged, it is just no longer in front of the response. This is the sibling of round 13's arm-publish finding; I fixed one path and left this one. P3 β€” enqueue was not fenced to the generation the capability decision was made against. The access checks, file resolution and owner re-read are all awaits, so the run can be replaced before the enqueue: the item then lands on the REPLACEMENT queue while the durable preempt flag and the arm still name the previous epoch, the arm is fenced out at the owner, and the 202 promises an interrupt that cannot happen. enqueueSteer now takes an expected generation, mirroring drain/peek, and the Redis path enforces it inside STEER_ENQUEUE_LUA so the check is atomic with the push rather than racing it. All three new specs verified counterfactually, including the Lua guard against real Redis (removing it returns 1 where -1 is required). * ⚑ feat: Interrupt & Steer β€” client half (PR 3 of 3) Makes preemptive steering reachable. Consumes the server contract from PR 2 (POST /chat/steer `preempt`, echoed on the 202) and the SDK seam in @librechat/agents 3.3.5. Settings shape follows the agreed correction, NOT the earlier plan draft: `steerInterruptsByDefault` is a boolean ORTHOGONAL to `duringRunDefaultAction` β€” that enum still chooses steer-vs-queue, the new boolean chooses how soon a steer lands. This deliberately avoids widening the enum to three values, which would have silently broken two hard-coded binary TOGGLES (`DuringRunAction.tsx`'s setter and `SteerMenu`'s `useDefaultToggleEntry`, both `prev === 'steer' ? … : …`) where a third value collapses to the wrong branch and one click erases the setting. - useSteering: `submitSteer` takes an opts bag and threads `preempt` into the POST, the optimistic chip, and the failure chip. The ACK relabels from the SERVER's echo, so a deployment that cannot seal mid-stream downgrades the chip's wording instead of erroring β€” the entire UX surface of capability degradation. New `interruptSteer` reuses the whole chip lifecycle and degradation ladder, and falls back to `interruptAndSend` when `!canSteer`, because steering needs a server-side job and an always-visible button would otherwise be dead for the whole first turn. `steerFromComposer` honours the new preference. - Composer: always-visible `InterruptSteerButton` with one fixed meaning (stop now, keep what's written), disabled on a paused run to pre-empt the server's 409, `type="button"` so it never steals the form's Enter submit, RTL-correct margins. A fourth hovercard row on the during-run send button, and ⌘/Ctrl+Shift+Enter routed AHEAD of the bare ⌘/Ctrl+Enter branch that would otherwise swallow it. - Chips: an in-flight preempt chip reads "Interrupting" with a ZapOff glyph; `preempt` survives reconnect through `seedSteerChips`. - `RunEnd.interruptArmed`, `drainAfterAbortByIndex`, `useQueueDrain`, `stopGenerating` and `interruptAndSend` are untouched β€” the preempt path deliberately shares none of the abort machinery. Tests: 7 new specs (posts preempt, turn-1 fallback, empty-text refusal, default route with and without the preference, server-echo relabel, double-click). 66 useSteering specs green; tsc and lint clean. Round-1 review fixes folded in: - P1: interrupt & steer no longer hard-aborts a run paused on tool approval. `canSteer` is false there, so the fallback was routing the keyboard and hovercard paths into `interruptAndSend` β€” discarding the partial answer, the exact opposite of what the action promises. The fallback is now scoped to the missing-conversation case only, and a paused run refuses outright (the standalone button was already disabled; the guard now lives where all three paths reach it). - The preference no longer leaks into the explicit Steer action. `steerFromComposer` backs both the default Enter route AND the explicit hovercard row / Ctrl+Enter alternate; applying `steerInterruptsByDefault` inside it made ordinary Steer interrupt and the two rows indistinguishable. It now takes an explicit argument that only `submitDuringRun`'s default route sets. - Retry preserves preemption: a failed interrupt-steer chip keeps `preempt: true`, and `retrySteer` now forwards it rather than silently resubmitting as an ordinary tool-boundary steer. - ⌘/Ctrl+Shift+Enter defers to a rebound submit shortcut, mirroring the bare ⌘/Ctrl+Enter branch β€” a user who bound submit to that chord keeps getting submit. Round-2 fix: the preempt label now survives the page-reload resume path too. `seedSteerChips` (useResumableSSE) and `restoreSteerChips` (useResumeOnLoad) are two independent TPendingSteerβ†’PendingSteer mappers with near-identical bodies; the first carried the flag and the second silently dropped it, so an armed interrupt reverted to plain "Steering" after a reload. Swept: those are the only two in production code. The reclaim/convert paths deliberately omit it β€” a queued follow-up starts its own turn, so there is nothing to interrupt. * fix: yield the interrupt-steer chord only to a submit shortcut bound to it The previous guard skipped the Ctrl/Cmd+Shift+Enter branch whenever ANY submitMessage override existed. Rebinding submit to something unrelated (Ctrl+J) or unbinding it entirely then fell through to the override resolver, which returns 'none' for shifted Enter β€” silently removing the shortcut the hovercard still advertises. Compare the pressed chord against the configured one instead. The adjacent bare Ctrl/Cmd+Enter branch keeps its any-override guard on purpose: that chord IS the default submit chord, so once submit moves the resolver should own it. The predicate already existed inside resolveSubmitOverrideAction; pulled it out as bindingsMatch so both sites compare chords the same way. That call is behavior-preserving β€” eventBinding.key is 'Enter' by the early return, and equal hashes imply equal keys, so the dropped explicit key check was redundant. * fix: disable the Interrupt & steer menu row while paused on approval interruptSteer hard-refuses when the run is paused for tool approval, but the hovercard row was never gated, so it rendered enabled and clicking it did nothing at all β€” no chip, no queue entry, no toast β€” at exactly the moment a user is trying to say "stop, don't run that command". The standalone button already gates on pausedOnApproval; the row contradicted it. Gated on pausedOnApproval rather than !canSteer like the steer row above, because canSteer is also false before a conversation exists, where interruptSteer deliberately falls back to interruptAndSend and the row must stay live for the whole first turn. Tests pin both directions and were verified counterfactually: removing the gate fails the paused case, and using !canSteer fails the first-turn case. * test: render the during-run hovercard eagerly instead of driving Ariakit The new spec passed locally and failed all four cases on CI's Ubuntu and Windows shards: Ariakit's show path keys off pointer geometry, which jsdom reports as zeros, so whether a synthetic mouseEnter opens the hovercard is environment-dependent. Driving it was testing Ariakit's hover behavior, not which rows this component disables. Mocking the three Ariakit primitives renders the rows unconditionally and drops the fake timers. Both counterfactuals still fail as they should: removing the gate fails the paused case, !canSteer fails the first-turn case. * test(e2e): cover interrupt & steer sealing mid-stream The mock Playwright suite covered every sibling during-run action β€” steer at a tool boundary, steer degrading to a queued follow-up, queue, and interrupt & send β€” but not interrupt & steer, the one this stack adds. Uses E2E_SLOW_REPLY, which streams pure text with no tools, so the scenario is the same one where an ordinary steer provably degrades to a queued follow-up turn. Injecting in-thread there is something only a mid-stream seal can do, which makes the assertion discriminating rather than incidental: the steer part lands in the response, the final chunk never arrives, the text written before the seal survives, and no follow-up turn pair is created. * test(e2e): assert the run resumes after the seal, not just that it sealed The other four assertions are all satisfied by a seal that killed the run: the steer part is persisted by applySteer during the drain, before the continuation starts, so 'sealed and resumed' and 'sealed and died' were indistinguishable β€” and resuming is the whole difference from interrupt & send. The continuation answers the injected steer, whose text carries no fake-model marker, so getLatestUserText falls through to the default reply. That string ('E2E mock reply') is distinct from the setup turn's ('E2E reply