I3389 — The mesh suite is run in both arms and its guarded verdict is published with its losses

✕BLOCKED wave 8 · p1 · implementation profile · parallel safe

Part of the-mesh-is-measured-against-one-agent — Coordinated workers on the mesh are compared with a single agent on the same batch tasks, at quality held by a registered non-inferiority guard.

Objective

Dry-run, rehearse one pair, then run every declared job of suite mesh in the registered order, commit the run records under .ai/repo/benchmarks/economics/runs/mesh, regenerate the reports and publish the T4 verdict with its guard, stating that it is a one-host result.

Why

T4 of the owner's report of 2026-10-19 asks for the speed-up to be shown; the result is published whether favourable or not.

Current state

No mesh run exists.

Desired state

runs/mesh holds every attempted run; T4 reads a computed status.

Scope

• .ai/repo/benchmarks/economics/runs
• docs/generated/economics.json
• docs/generated/economics.md
• docs/generated/economics.yaml
• site/data/generated/economics.json

Out of scope

• Editing the registration during the runs
• Excluding a run

Dependencies

• I3388✕BLOCKED The mesh suite is registered before its first run, and the T4 protocol names its registration digest
• I3354✕BLOCKED A pure scorecard function computes each proof test's status, counted sample, effect and losses from committed records, and withholds a verdict until the registered sample is complete

Acceptance criteria

• Every declared pair is attempted, and each missing or invalid pair is named
• The guard's bound is published beside the verdict
• The published result states it is a one-host result
• majordomus economics check and majordomus generate --check pass

Validation

• majordomus economics summary
• majordomus economics proof
• majordomus economics check
• majordomus generate --check

Evidence required

• mesh_run_records
• mesh_verdict_published

Evidence

None recorded. Every token above needs a command or an artifact behind it before this issue can be completed; narrative is refused.

Timeline

started

—

verified

—

completed

—

Those three fields, the evidence above and the state of the dependencies are all the status is made of. There is no status field to disagree with them.

Canonical record: .ai/repo/project/issues/I3389.yaml. Read it back with majordomus plan show I3389.