info-tech-canon/workplans/INFO-WP-0019-conformance-and-maintenance.md

---
id: INFO-WP-0019
type: workplan
title: "Conformance, reproducible consumption, and maintenance"
domain: infotech
repo: info-tech-canon
status: finished
flavor: implementation
owner: codex
topic_slug: canon-optimization
created: "2026-09-05"
updated: "2026-09-09"
state_hub_workstream_id: "b29261ba-c1e4-5533-8185-ab2d5b433685"
---

# Conformance, reproducible consumption, and maintenance

Implements the optimization recommendations in
`history/2026-09-05_002437+0200-scope-against-intent.md`.
Reviewed against the current working tree; authorized by the operator on
2026-09-05. Uses the AGENTS.md INFO-WP file convention.

## Reusable conformance and coverage

```task
id: INFO-WP-0019-T01
status: done
priority: high
state_hub_task_id: "abc0ed79-009f-5254-a2bf-4dcedd1925e9"
```

Extract shared schema validation and emission-cadence semantics into contract
modules. Expose arbitrary declaration validation and explicit coverage through
the CLI. Reject unsupported versions, malformed declarations, and duplicate
source IDs. Enforce explicit ownership conflicts. Preserve all existing tests.

## Reproducible contract bundle and adoption gate

```task
id: INFO-WP-0019-T02
status: done
priority: high
state_hub_task_id: "1ec85dc5-5775-50b6-8066-ac839fd252c9"
```

Provide deterministic, content-addressed export of the cadence contract,
schema, and example; refuse overwriting different contents. Define owner-authored
adoption evidence requirements and keep draft promotion contingent on two
independent implementations. Local examples must not count as adoption evidence.

## Freshness and evidence

```task
id: INFO-WP-0019-T03
status: done
priority: high
state_hub_task_id: "e4171772-50bf-58f6-9c87-8eddbef1caa5"
```

Add non-mutating generated-asset checks, reproducible scope inventory, and
revision/time/source-digest evidence on validation reports. Refresh work-record
projections via State Hub. Check stale and missing generated assets with tests.

## Portable consumption

```task
id: INFO-WP-0019-T04
status: done
priority: medium
state_hub_task_id: "e258d84d-d969-5032-baef-ae1060843319"
```

Remove hardcoded workstation dependency paths. Resolve an installed bench
distribution and an explicit external corpus root. Prove wheel installation
and commands from an unrelated directory. Document upstream packaging limits
without changing sibling repositories.

## Measure performance and verify implementation

```task
id: INFO-WP-0019-T05
status: done
priority: medium
state_hub_task_id: "09a585e8-4c67-5143-bc3a-ae47623547ea"
```

Measure repeated reads before deciding whether caching is justified. Run all
tests, canon/profile validation, generated freshness checks, and the clean
installation proof. Update SCOPE/README and record evidence. Avoid speculative
models or runtime expansion.

## Independent consumer adoption

```task
id: INFO-WP-0019-T06
status: done
priority: high
state_hub_task_id: "b33aeb8c-ba7c-5094-9902-85565c1e721b"
```

Requires two source owners to provide independently implemented declarations
and consumer results pinned to the contract digest. This repo can validate
submitted evidence but cannot manufacture source ownership or deployed proof.
No external messages or consumer changes are authorized by this workplan.
Promotion remains blocked until that evidence arrives.

2026-09-09 recheck: the inbox contains an InterfaceCanon semantic-boundary
review request, and `feedback/` still contains only the two resource-control
capability reports. Neither supplies independent emission-cadence declarations
or the required pinned consumer results. T06 remains waiting; boundary review
is handled separately in INFO-WP-0026.

2026-09-20 recheck: no change in the evidence, and the gate is verified ready.

`feedback/` still holds only the two 2026-08-15 resource-control capability
reports. `incoming/` holds only its README. The inbox carries eleven messages,
the most recent a 2026-06-27 concept contribution from ops-warden; none is an
emission-cadence declaration. The only declaration in the repository remains
`standards/emission-cadence/examples/qonto-assistant.yaml`, which the standard
itself states is not the source-owned declaration and makes no claim about live
emission.

What this repository owes the gate is in place and was re-tested rather than
assumed. `emission-review` validates an arbitrary source-owned declaration
against the named schema and reports `operational_truth_assessed: false`, so a
submission cannot be mistaken for observed behaviour.
`export-emission-contract` produces a content-addressed bundle whose manifest
carries the gate in words — "Two independent source-owned implementations
required; example does not count" — at digest `972c0b6701d1693f`. A source owner
can pin to that digest today.

The blocker is therefore not readiness. It is that no external source owner has
been asked. The contract came from the King's Guard taxonomy and the evidence
classes remain owned by NetKingdom, which makes `netkingdom` the obvious first
source owner; a second must be independent of it. Asking them is an outward
action this workplan does not authorize, so T06 stays `wait` until the operator
decides whether to solicit declarations, and from whom.

### Result — 2026-09-21 (T06)

Closed by restating the gate rather than by evidence arriving, on the operator's
decision of 2026-09-21 after the third recheck found the evidence unchanged.
Recorded as INFO-DEC-2026-002, with the reasoning in
`history/2026-09-21_000846+0200-emission-cadence-promotion-gate.md`.

The gate as written — "two independent Emission Cadence source implementations"
before draft promotion — made one sentence do two jobs: whether the contract is
specified well enough to implement against, which this repository can answer,
and whether it survives implementers who did not write it, which it cannot. The
second blocked the first, so the standard sat at `draft`, a word that reads as
"unfinished specification", when what was missing was adoption.

The gate is now two levels, written into the standard as section 10 rather than
only into SCOPE.md. **Candidate** requires schema coverage of every form,
one-command validation of an arbitrary source-owned declaration, a
content-addressed bundle a source can pin to, separation of declaration from
observation, and a documented external consumption path — all met and re-tested.
**Stable** requires two source-owned declarations from independent owners
pinned to a digest, an observer result from a third party, and a recorded
incompatibility or recorded absence of one.

Emission Cadence moves to `candidate` at document version `0.2.0`. The wire
contract is untouched at schema version `0.1`; a producer conformant yesterday
is conformant today.

What was given up is stated in the decision record and worth repeating here:
`candidate` now carries no interoperability evidence at all. The mitigation is
that the standard says so in its own section 10, naming that no consumer has
implemented the contract and that its worked example is not a source-owned
declaration. The stable gate is stricter than the sentence it replaces, since
it adds the observer result and the incompatibility record.

## Implementation notes

T01–T04 implemented in `contracts.py`, `maintenance.py`, `paths.py`, CLI,
service integration, and `docs/canon-consumption.md`. Explicit mapping schema
bindings and ownership checks improve coverage without claiming exhaustive
semantic validation. The normal bench initializer imports unrelated database
integrations; installed-module discovery retains the narrow adapter. Clean
wheel consumption uses the documented explicit-dependency installation until
the upstream owner separates reference-data packaging.

Registration: State Hub workstream `b29261ba-c1e4-5533-8185-ab2d5b433685`.
The consistency tool reports an INFO-WP/ITC-WP prefix advisory: this plan follows
the explicit INFO-WP convention in AGENTS.md; existing plans and the registry
use ITC-WP. This naming inconsistency does not prevent registration.

## Verification — 2026-09-05

`make check`: 36 tests passed (22.42 seconds); all 77 generated assets current;
canon validation passed (76 artifacts, 315 relationships); Small-SaaS passed
(15 payloads). Installed-wheel inspect, validation, and emission review passed
from an unrelated directory with PYTHONPATH unset and a copied corpus.
Ten uncached inspections measured median 788.05 ms, maximum 1241.92 ms;
caching remains deferred pending a workload/latency requirement.

Local implementation T01–T05 is complete. Overall status is blocked only on
T06's independent owner evidence. See
`history/2026-09-05_optimization-implementation.md` for artifacts and limitations.