| Subject | Status | Report date | Age | Note |
|---|---|---|---|---|
| checkout | assessed | 2026-08-05 | 9d | — |
| search | assessed | 2026-08-05 | 9d | — |
| inventory | assessed | 2026-08-06 | 8d | — |
| pricing | assessed | 2026-08-07 | 7d | older copy of the shared ancestor |
| fulfilment | assessed | 2026-08-07 | 7d | own harness, no shared lineage |
| legacy-batch | assessed | 2026-08-08 | 6d | declared testing variance |
6 of 6 subjects assessed; 0 stale; 0 unreachable.
| Subject | AB | AI | WF | OM | TP | OR | OA | HR | WP | AC | AD | Te | Ob | Gv |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| checkout | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 2 | 3 | 3 | 2 | 3 |
| search | 3 | 2 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 2 | 3 | 3 | 2 | 3 |
| inventory | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 2 | 2 | 3 | 3 | 2 | 3 |
| pricing | 2 | 2 | 3 | 2 | 3 | 2 | 3 | 2 | 2 | 2 | 2 | 2 | 2 | 2 |
| fulfilment | 3 | 3 | 4 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 4 | 3 | 3 |
| legacy-batch (maintenance) | 2 | 1 | 1 | 2 | 1 | 2 | 2 | 2 | 2 | 1 | 2 | 1 | 1 | 1 |
Level in cell, coloured L1–L5. ᶦ inherited ·
ᵘ inherited-unbound · unmarked is local. Red outline marks the
weakest cell in the column among active subjects. Non-active rows are
de-emphasised and do not set the ceiling.
AB Agent behaviour · AI Agent input · WF Workflow · OM Operating model · TP Teams provide · OR Output role · OA Output artefact · HR Humans review · WP Work patterns · AC Agent composition · AD Agents… · Te Testing · Ob Observability · Gv Governance
| Subject | Habitat maturity mean | Cognitive level | Gap | Regime |
|---|---|---|---|---|
| fulfilment | 3.14 | 3 | -0.14 | Coherent |
| checkout | 2.86 | 3 | +0.14 | Coherent |
| search | 2.79 | 3 | +0.21 | Coherent |
| inventory | 2.79 | 3 | +0.21 | Coherent |
| pricing | 2.21 | 3 | +0.79 | Ambition outpaces enablement |
| legacy-batch | 1.50 | 3 | +1.50 | Ambition outpaces enablement |
Four of the five active services cluster tightly; pricing is the outlier. Including legacy-batch the range is 1.64, but it is marked maintenance and does not set the agenda.
HARNESS.md descended from a common ancestor — checkout, search, inventory, pricing — same section structure and rule names, diverged bodies.Agent composition, Observability — weak in four of five active subjects. Neither appears in any copied harness, so every team was left to invent them and four did not.
pricing is behind across the board, and the lineage analysis says why: it runs an older copy of the shared ancestor. legacy-batch is excluded from common-weak by its maintenance posture — otherwise it would set the ceiling on eleven of fourteen dimensions.
Extract the common layer — and treat the divergences as the work. Four services maintain four copies of one harness. The extraction is not the hard part; reconciling what the copies disagree about is. Three Python versions and two coverage thresholds encode four local decisions — some real requirements, some accidents nobody remembers making. Go through them deliberately rather than picking the newest copy and declaring it canonical.
Start with pricing. Widest active gap, cheapest to close: it is not missing capability, it is missing the CI-gate section the other three copies already have.
Put agent composition and observability into the extracted layer, not into each service — they are weak in four of five precisely because no copy of the ancestor ever had them.
Leave fulfilment alone for now. It shares no lineage, it is the only subject whose habitat is ahead of the team, and folding it in would mean reconciling a fifth unrelated design for no benefit.
Declared variance: legacy-batch · Testing · “COBOL batch; harness test tooling does not apply” — listed as declared, not reported as drift.