Review a long record within a bound sized to it, and let a proposal waiting on its author leave the knowledge base healthy (hq issue 508)

A judgement over one bus call's 30 seconds is collected by its ticket from
claude-code, so an eleven-clause record can be reviewed; a proposal out of
review tries is told to the operator, not raised as the module's fault, so
it no longer holds the module's own rollouts.

modules/mesh-knowledge/README.md

@@ -99,7 +99,10 @@ as text. `get` serves each, and any other field a record of hq wrote under `fiel
    the record, the records it links and the rules, and answers `approve` or `changes` with findings. The rules it sees
    name every kind of link `propose` accepts, from the one list `propose` holds a link to, and the record it sees
    shows its state in the knowledge base apart from a research effort's own status (hq issue 492). The background
-   work asks it at once, before the next step of any drafting, one judgement at a time.
+   work asks it at once, before the next step of any drafting, one judgement at a time. Every judgement asks a bound
+   sized to its question: 25 seconds and 2 more per KiB, at most 120 (hq issue 508: an eleven-clause record was never
+   reviewed in 25). A judgement not done within one bus call is answered pending with a ticket, and collected from
+   the same node with `claude-code.claude_code_judged`.
    - On `changes` the record stays proposed, the findings kept with the version they judged (`get` shows them).
      Its author revises it by proposing it again, and that version is reviewed again.
    - A judgement that cannot be had (the judge unreachable, or an answer that is not `approve`, or `changes` naming
@@ -108,8 +111,9 @@ as text. `get` serves each, and any other field a record of hq wrote under `fiel
      version in the store and said in its history, so the background work takes it up at that time, also after a
      restart of the tool runner (hq issue 497). After 3 tries the record says it was the last and its author
      proposes it again to have it reviewed again; the operator is told once on the operator's channel (keyed by
-     record and version, and cleared once the record is proposed again, accepted, withdrawn or superseded), and
-     the module is unhealthy, naming it, so the mesh raises a condition.
+     record and version, cleared once the record is proposed again, accepted, withdrawn or superseded, and sent
+     again at the next step when it could not be sent). The module's health names the record and stays healthy:
+     the record waits on its author, not on the module, so it never holds the module's own rollouts (hq issue 508).
      An ask of the operator that cannot be made counts as a try the same way.
    - A review whose outcome the database refused to keep is held back for that record alone (10, then 20
      minutes, 3 tries, logged as given up), so it never blocks the records behind it.