fix(tasks): correct the change-feed contract and its documentation

Review found the `onTasksChanged` teardown reasoning inverted. The comment
claimed every registration is an effect on the registry's own fiber, so
listeners would be gone before service disposal empties the store — but the
traceable proxy rebinds `this.ctx` to the CALLER, which this package's own
HMR-safety test already proves. The only shipped consumer registers from the
api-proxy mux stream, so it was still listening and simply kept the rows it
last received. Service disposal now announces the emptied set, and teardown
announces its stopping transition immediately instead of leaving an observer
on `running` for however long a slow producer takes to release.

Two documentation claims were false in the opposite direction: the Agent Note
and the ui-task README both said an unowned task is invisible in the header,
while `list(caller)` returns unowned tasks to every caller, the carrier fans
their changes out to every subscribed session, and this PR's own test asserts
exactly that. The note even contradicted itself two sections earlier. Both
sides now state the real asymmetries — another session's tasks, and the
process-local registry emptying on restart.

The "no Web path calls the consuming `ctx.tasks.read()`" invariant claimed a
test that did not exist; the carrier suite's producer had no `readOutput` at
all, so a stray read would have failed nothing. Its producer now counts cursor
consumption and the lifecycle and baseline paths both assert zero.

Also: a session created after the mux opened now receives the task baseline it
missed, the popover samples its clock when it opens rather than at mount, and
a failed task's unbounded producer detail elides instead of widening the row.

packages/client/ui-task/README.md

@@ -21,4 +21,4 @@ None; the package never assembles or sends provider requests.
 ## Known Limitations and Deferred Work
 
 - **Rows are read-only** — a task's streamed output and a human-initiated cancellation are separate phases. Cancellation additionally owes a model-facing decision the seam does not answer today: `kill()` marks terminal delivery reported, so an interrupt written against the current contract would leave the model believing its task is still running.
-- **The list is not the registry's own set** — it shows what the owning session can see through the wire view, so an unowned task (one started without a live `Agent`) is invisible here while `task_list` still reports it to the model.
+- **The list is not the registry's own set** — it shows what one session can see through the wire view, so a task owned by another session never appears here, and a process restart empties the list while the transcript keeps the `run_in_background` cards that started those tasks. An unowned task (one started without a live `Agent`) is the opposite case: it reaches every session's list, matching what `list(caller)` reports to every caller.