fix(schema): restore constraints canonical-forecast-point dropped

[bokelley]

Summary

Implements item 7 of #6628. The canonical twin of forecast-point.json silently dropped three normative constraint groups its source carries:

• the maximum: 1 bounds on viewable_rate range values,
• the maximum: 1 bounds on metrics.coverage_rate range values,
• the anyOf requiring standard whenever any viewability value is present.

Net effect before this fix: a canonical payload could carry viewable_rate: {mid: 1.3} under no stated viewability standard — a shape the source schema rejects — because nothing compares canonical pairs against their source twins (the enum-drift linter only covers inline enums).

This restores the constraints in the canonical file's compact style and adds tests/canonical-forecast-point-parity.test.cjs, which asserts the twins' viewability property sets match (with the brand-ref→brand-key swap pinned as the one intended difference), the anyOf is identical, the rate bounds agree, and — behaviorally, via Ajv — the canonical twin now rejects out-of-bounds rates and standard-less viewability values while accepting well-formed rows.

Constraint tightening on the unreleased canonical 3.2 surface; minor changeset. Item 8 of #6628 (the reporting-webhook frequency subset) remains WG-queued — deliberately not bundled here.

🤖 Generated with Claude Code

[aao-secretariat]

Ladon verdict: Escalate to human review

Escalate — gated path requires human/CODEOWNERS approval.

This PR modifies static/schemas/source/core/canonical-forecast-point.json, which matches the repo's gated-paths glob static/schemas/source/** (gated_paths: true). The required review status is REVIEW_REQUIRED, not APPROVED, so the hard approval gate (decision row 2) is not satisfied and the PR cannot be auto-approved.

The reviewer found no blocking issues: the change restores three normative constraint groups (viewable_rate/coverage_rate maximum: 1 bounds and the standard-required anyOf) to align the canonical twin with forecast-point.json's source, in the compact allOf style. The restored anyOf is byte-for-byte structurally identical, rate bounds match, a new Ajv-backed parity test pins all three groups plus behavioral accept/reject, and the change is correctly carried by a minor changeset with no schema↔docs drift. This is constraint tightening on the unreleased canonical 3.2 surface — a Normative (non-breaking) change, so the breaking-class hard gate does not independently apply.

The diff is clean, but the schema source file is under a deterministic hard approval gate that only lifts when review_decision is APPROVED. Human/CODEOWNERS review is required before merge.

Escalation reasons

• Modifies static/schemas/source/core/canonical-forecast-point.json (gated path, static/schemas/source/**); review_decision is REVIEW_REQUIRED, not APPROVED.

Why human review

• Modifies gated path static/schemas/source/core/canonical-forecast-point.json (static/schemas/source/**) while review_decision is REVIEW_REQUIRED, not APPROVED — human/CODEOWNERS approval required (decision row 2).
• This PR touches a path under a hard, non-overridable approval gate (static/schemas/source/core/canonical-forecast-point.json (modified) matches static/schemas/source/**) and the current GitHub review decision is 'REVIEW_REQUIRED', not APPROVED. This is a hard gate enforced in code — Ladon cannot auto-approve until a human/CODEOWNERS approval is recorded, regardless of how clean the diff is.