Skip to content

Example: parent repo with submodules

Synthetic example — constructed to illustrate the report shape.

These numbers are invented to show what the report looks like when a shared harness binds everywhere. They are not a reading of any real estate. The self-assessment is the real one.

One HARNESS.md at the root of a parent repository, three application submodules underneath, all bound to it.

The manifest

version: 1
team: platform-group

habitats:
  - id: root-harness
    kind: self
    path: .
    provides: [HARNESS.md, AGENTS.md, ci]

subjects:
  - id: web
    path: ./web
    habitat: root-harness
  - id: api
    path: ./api
    habitat: root-harness
  - id: workers
    path: ./workers
    habitat: root-harness

Coverage ledger

Subject Status Report date Age Note
web assessed 2026-08-04 10d
api assessed 2026-08-04 10d
workers assessed 2026-08-05 9d

3 of 3 subjects assessed; 0 stale; 0 unreachable.

Dimension matrix

inherited · inherited-unbound · unmarked is local · * weakest in column

Subject AB AI WF OM TP OR OA HR WP AC AD Te Ob Gv
web 3 3 4ᶦ 3 4ᶦ 3 3ᶦ 3 3 3 3 3 2* 4ᶦ
api 3 3 4ᶦ 3 4ᶦ 3 3ᶦ 3 3 3 3 3 3 4ᶦ
workers 3 2* 4ᶦ 3 4ᶦ 3 3ᶦ 3 2* 2* 3 2* 2* 4ᶦ

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

Five dimensions are inherited in every subject, and none are inherited-unbound. The root harness reaches all three submodules — each subject's CI reuses the root workflows and each AGENTS.md defers to the root file rather than restating it.

Gap table

Subject Habitat maturity mean Cognitive level Gap Regime
web 3.14 3 −0.14 Coherent
api 3.21 3 −0.21 Coherent
workers 2.86 3 +0.14 Coherent

Spread

−0.21 (api) to +0.14 (workers) — range 0.35.

A narrow spread is the signature of a habitat that genuinely reaches everything it claims to. The same team, working the same way, meets substantially the same environment in all three submodules. What differs between the extremes is small: api has slightly more observability wired up; workers has thinner test and composition coverage.

Split ceiling

Common weak — Observability. Weak in two of three subjects. This is the enablement backlog: the root harness does not supply observability scaffolding, so each submodule is left to invent it. Fixing it once at the root fixes it three times.

Locally weak — workers: Agent input, Work patterns, Agent composition, Testing. These are that team's own backlog and do not generalise.

Confidence

42 placements feed the matrix: 24 observed, 18 inferred, 0 asked. No subject exceeds five inferred dimensions.

What this estate should do next

Lift observability at the root, not in each submodule. It is the one dimension weak across the estate, and the binding already works — a change to the root harness propagates without asking three teams to do anything.

Leave the workers gaps to the workers team. They are real but local, and pulling them into an estate-wide programme would slow the fix and blur who owns it.

Do not chase the levels. All three subjects are Coherent; nothing here is out of step with itself.


Portfolio regime naming — Federated, Distributed, Fragmented, Islanded — arrives in a later release. This example shows the shape a federated estate produces, without asserting the label.

View this report as a shareable HTML page — a single self-contained file, no network assets, works opened straight from disk.