I3429 — One entity reads the same on the command line, REST, MCP and the Cockpit: identity, kind, class, evidence, observation envelope and edge order

✕BLOCKED wave 2 · p2 · implementation profile · parallel safe

Part of an-entity-says-how-it-is-known-and-what-it-affects — One entity read names how its record says it is known and when it was observed, identically on every surface, and what a change affects is derived from the same index.

Objective

Extend the Cockpit integration test with one test that reads four surfaces for one entity of every kind the index holds: majordomus entity show as JSON, GET /api/v1/entity, majordomus_entity over MCP and the Cockpit entity page model. It compares uri, id, kind, epistemic class, evidence state, the observed envelope and the order of relations.

Why

The owner's 2026-10-19 Cognitive Fabric pack requires cross-surface tests asserting the same identity, classification, deterministic order and freshness semantics (campaigns/prismatic-way-cognitive-fabric-prompt-pack/prompts/05_SURFACES.md, and the verification/ACCEPTANCE.md row on CLI, API, MCP and UI).

Current state

tests/properties.rs compares MCP with HTTP for GET routes only, and never the command line (apps/majordomus-cli/tests/properties.rs:290-300). tests/cockpit.rs compares the Cockpit entity page with GET /api/v1/entity (apps/majordomus-cli/tests/cockpit.rs:187-245). No test covers entity.show on all four surfaces.

Desired state

A divergence in any compared field on any surface fails one named test.

Scope

• apps/majordomus-cli/tests/cockpit.rs

Out of scope

• Extending the registry-wide property to writers (milestone surfaces-never-disagree)
• Comparing rendered HTML text

Dependencies

• I3427✕BLOCKED Every entity read names how its record says it is known, and never answers verified itself: proof stays the pointer it is
• I3428✕BLOCKED An entity read carries the observation envelope repository.info already answers, plus the blob of its file, and a change to that file moves both

Acceptance criteria

• The test covers one entity of every kind entity.kinds reports
• Changing the class or the relation order on one surface in a fixture fails it, naming the surface and the field
• It runs without network and without the checkout's .ai/local

Validation

• cargo test --manifest-path apps/majordomus-cli/Cargo.toml --test cockpit

Evidence required

• entity_surfaces_agree

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/I3429.yaml. Read it back with majordomus plan show I3429.