fix(sync): finalize tool parts orphaned by an interrupted turn after settlement

When a managed OpenCode process dies mid-turn (crash, health-check
restart), the persisted turn never settles: the trailing assistant
message has no time.completed and its tool parts stay pending/running
forever — the server never finalizes them (anomalyco/opencode#19023).
The existing settle-triggered tail refresh refetches the same stale
records, so the UI kept running tool timers and working styling
indefinitely (#2577).

Now, when a session is authoritatively settled (session.idle/
session.error event, or an authoritative status snapshot lowering a
previously busy session) and the trailing assistant message is still
unfinished with active tool parts and no pending question/permission,
the orphaned parts are finalized locally as error/"Interrupted" with
an end time — the same shape OpenCode itself writes for cancelled
tools. The mark is gated on an explicit idle status (absent status is
"unknown", never judged), never applies while busy (including
question/permission waits), and a later terminal event or refresh
supersedes it while a stale running refresh cannot regress it (the
reducer and materializer already preserve final statuses).

Fixes #2577

packages/ui/src/sync/DOCUMENTATION.md

@@ -204,6 +204,8 @@ Incomplete-session materialization is deduplicated by runtime, directory, and se
 
 When `session.idle` or `session.error` settles a session but the trailing assistant message still contains a `pending` or `running` tool, sync refreshes that session tail. This narrowly reconciles a missed terminal tool-part event without refetching normally completed turns or stale tools from older turns. A stale refresh or delayed part event cannot regress a locally observed terminal tool to an active status.
 
+When a session is authoritatively settled — `session.idle`/`session.error` event, or an authoritative status snapshot that lowers a previously busy session — and the trailing assistant message is still *unfinished* (`time.completed` missing) with active tool parts and no pending question/permission, the turn is treated as interrupted (managed OpenCode process died mid-turn; the server never finalizes the parts, see openchamber#2577 / anomalyco/opencode#19023). The active parts are finalized locally as `error`/`Interrupted` with an end time, so tool timers stop and cards render the error state. The mark is gated on an explicit idle status (absent status is "unknown", never judged), never applies while the session is busy (including question/permission waits), and a later terminal event or refresh supersedes it while a stale `running` refresh cannot regress it.
+
 Directory stores also own session-keyed sidecar notification channels for permissions and message materialization. High-frequency realtime part events annotate the exact session/message before committing, so visible records, user history, renderability, and sidebar permission rows are not notified by unrelated sessions. Structural message replacements notify only changed subscribed session buckets; unannotated bulk part replacement conservatively resets active message subscribers so bootstrap, pagination, rollback, and legacy writers cannot leave stale projections.
 
 Message sidecar consumers also filter targeted updates by purpose before notifying React. Suspended live-tail text/reasoning changes do not rebuild visible message records, but structural Task session identity changes bypass suspension so a parent can link a newly created subagent immediately. Assistant-only part changes do not rebuild user input history, and targeted updates that preserve authoritative part buckets do not recheck a session that is already renderable. Message replacements, removed final part buckets, and conservative resets always notify.