The parked composition-at-running path is folded as the lead's ruling Q3: one
path, so no check composes with the running engines' own code (ADR 0656 §5).
The sandbox request no longer drops a change's manifest verdict: it carries the
check's judge's, or none, which the composer reads as accepted (ADR 0578 §3).
Claude-Session: https://claude.ai/code/session_01NcbgfFSnn24M3GcqttoU7i
B0's shape wins: the running checker asks and records facts, the checking program
decides. The registry's types win: the rules and the program read the generated
checker contract and the domain SDK, and no mesh-sdk delivery import is left.
The registry's series, renames and unparsed seat definitions become facts with
their refusals as rules; the tagged versions ride on the case. The program reads
the heads' seat definitions through sdk/seatdefs and the running checker lays them
over its reader's list by seat, reading no file of the heads: the reader's list
is larger than a run input's bound.
Claude-Session: https://claude.ai/code/session_01NcbgfFSnn24M3GcqttoU7i
The running checker asks the checking program twice on the build seat's job run: first the
heads' seat list, which the manifest validator's requests (change and trunk) and the
node-engine composer's carry; then the verdict of the facts it gathered. The program is built
at the group's mesh-delivery head, else at mesh-delivery's trunk, and its input is kept in the
checker's new state inputs as the exact bytes of its digest, read with checker_inputs. A run
ask no build node begins within 5 minutes is put once more, then fails the check naming the
job; a program that does not build or answer fails the check naming it, and the verdict
records who judged it (JudgedBy). A second step-seen hook on a context is added to the
first, so a try still sees its steps.
Claude-Session: https://claude.ai/code/session_01NcbgfFSnn24M3GcqttoU7i
The group try of 2026-11-06 10:07 composed with a composer that had no
seat list, so its composition failed on all four machines for paths and
holders the seats define. The checker's one list per check (the
reader's, the members' heads' laid over it, ADR 0636 §4) now goes to the
judge and the sandbox composer alike; a list the reader has not read
goes to neither, and the composer then judges no machine (ADR 0636 §1).
Claude-Session: https://claude.ai/code/session_01NcbgfFSnn24M3GcqttoU7i
The seat verb try gains members (ADR 0637 section 6); the try's group name and the missing-preview reason are one definition of mesh-delivery-sdk (ADR 0637 section 8).
Claude-Session: https://claude.ai/code/session_01NcbgfFSnn24M3GcqttoU7i
History (A2j finding 1, J4): the indexer's and the assigner's read tools are views.ReadTool over views.KeptSource, which
answers `{key, history: true}` now; modules (20), heard (64) and wants (20) keep more than one revision. Rows in the
view's "each tool answers through the owners' read tools as before": indexer_module_history and indexer_heard answer
the owner's kept history, and a want's history (view_tiers' Since) is read through assigner_read_wants (the row it
replaces held that the owner answered no history).
Series (A2j finding 2, J4): the checker keeps each series change as a delivery.SeriesChange {module, artifact, from,
to} and renders the status line from it; delivery.Verdict.Series holds the case's changes (each module's artifact once),
and the verdict event's series the pull request's (meshdeliverychecker.VerdictHolderEventSeriesItem). An artifact new
on the head is no series change (the lead's ruling). Rows: TestTheIndexersJudgeJudgesTheChangesManifests (the verdict's
series; failed first, empty) and TestTheStatusIsPendingWhileTheCheckRunsAndEndsSuccessOrFailure (the event's; failed
with the fill cut).
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
Design 32 (A package, its series): the review of a change sees every series it raises. For each module a
pull request touches, every artifact its trunk manifest and its head's manifest both declare whose series
differs is named as "<module>/<artifact> series <trunk> → <head>" in the pull request's mesh/check line,
never dropped when the line is cut; a manifest that cannot be read is said in the same place. Rows in
TestTheIndexersJudgeJudgesTheChangesManifests (raised: named; kept: not named); the raised row failed
first. gofmt of the integration test's imports (M2).
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
The forge carries the operator's own repositories (jschoubben/hal, jschoubben/novadock) beside the mesh's; the
indexer writes their pull requests as candidates, the gate never names their trunk, so the checker said each
"cannot be judged" and owed the status error on mesh/check, which the forge module wrote into repositories the
mesh's forge account may not write: 403, 119 times each (sweep 2026-10-30 item 7).
Quick fix: name the two repositories as not the mesh's in the checker.
Thorough fix, chosen because any repository on the forge can meet it: a candidate every pull request of which is
of a repository the gate's source.main names no trunk of and touches no module is outside the mesh. The checker
writes it no verdict and owes it no status, takes away a status owed before, says it once in the log, and lists it
in checker_status as outside_the_mesh. A new mesh repository's first pull request adds a module, so it is still
said untrunked and owed error (issue 1422). Nothing is outside before the gate's source is read.
A repository never merged through the gate has no entry in source.main, so its pull requests were dropped without a log
line, a verdict or a status, and checker_status counted them as new. Now the verdict says it cannot be judged and why,
the forge status is error with those words, checker_status names it under cannot_be_judged, the log says it once per
head, and the heads are judged whole once the trunk head is known. Quick fix: this. Thorough fix (the gate seeding
source.main from the pull message's base head) is still to do: it needs mesh-sdk, the forge bridge and the gate.
The checker wrote a try's asks only at its end, so a 31-minute try could not be told queued from
running (issue 1890). The try's record now names each ask as it is put, then the node that took
it, its state and the phase its step says, written on the step key's own changes. build_usage
names a running try's asks and takes a build module's capacity.<node> (configured bounded by its
node's build CPU, issue 1806's follow-up) when it is current with mesh-config.
The operator's question (how much are the build nodes used, is ace eligible) took six calls over four tools.
build_usage reads it from the bus's states: seats held (have), capacity (mesh-config resolved), pauses, the
build modules' steps by kind with left-running keys, the last hour and day with total and median time and time
at capacity, the running tries, and the backlog the view program reads from the seats' streams. A try now names
the asks it put, and the checker announces its tries readable, so the view tells a try's asks from a check's.
mesh-controller's try asked build-agent, which leaves the checking at R4e. A try is now a key in the checker's state
tries: the checker judges the head merged onto its trunk as it judges a pull request's head, through the same
judge, places and holds, writes its end in the key and owes the forge nothing (no status, no preview, no verdict).
One try per head, at most four at once; hq and a repository whose trunk the gate does not name are refused with
"no level of <repo> is checked by the checker".
The gate decides on mesh/check alone and matches an Approve's members by the heads hash over lowercase repositories
and heads. A success judged on an old trunk stood on the forge while the next check waited for a place or a delivery,
and a candidate whose repository the forge named in capitals carried a hash the gate never finds. A check that ends
after its trunk moved now owes pending, not its verdict.
jschoubben