I3368 — Six large tasks join the economics corpus, each failing on its start, passing on its reference and reviewed by someone other than its author

✕BLOCKED wave 3 · p1 · implementation profile · parallel safe

Part of the-corpus-is-registered-before-it-runs — A corpus of 20 tasks, stratified by a difficulty proxy measured and registered before the new tasks are written, is registered with its analysis, run order, stopping rule and a pinned release.

Objective

Add six large tasks on the ledgerlite fixture, within the registered large cut-offs: credit-note-lifecycle, multi-currency-report, storage-v3-migration, partial-payments (two sessions), vat-reverse-charge and csv-batch-import, each with task.yaml, prompt, hidden acceptance test and reference patch in the existing layout. Ask a review of the prompts and tests through majordomus_mesh_review_request from a reviewer who did not write them, and record the answered review ids in the task declarations.

Why

T2 of the owner's report of 2026-10-19 asks for 20 comparable tasks split by difficulty. The corpus has no large task. The critique asked for review by a party that is not the treatment's author; a peer review answered on the mesh is the available independent check, and its id is recorded.

Current state

grep complexity .ai/repo/benchmarks/economics/tasks/*/task.yaml gives 4 small, 3 medium, 0 large; the fixture holds 9 modules under test/fixtures/economics/ledgerlite/ledgerlite.

Desired state

Six new large tasks proven by economics references and reviewed.

Scope

• .ai/repo/benchmarks/economics/tasks
• test/fixtures/economics/acceptance
• test/fixtures/economics/reference

Out of scope

• Editing the fixture or an existing task
• Adding the tasks to the pilot suite

Dependencies

• I3374✕BLOCKED Task difficulty is a proxy measured from each task's reference, and the stratum cut-offs are registered before any new task is written

What waits on this

• I3370✕BLOCKED The corpus suite is registered and committed before its first run, and the T2 protocol names its registration digest

Acceptance criteria

• Each task declares complexity large and its proxy falls inside the registered large cut-offs
• economics references reports each failing on its start and passing on its reference
• No prompt names the hidden tests or the reference
• Each acceptance test uses only the Python standard library
• Each task names an answered review id from a reviewer who did not write it

Validation

• majordomus economics references --work-dir \"$(mktemp -d)\"
• bash test/run.sh 430_economics
• majordomus plan validate

Evidence required

• references_large_stratum

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/I3368.yaml. Read it back with majordomus plan show I3368.