TASK-IMP-089: task@1 template drops the duplicate out-of-scope section 4

Summary

Every conforming spec has carried two out-of-scope homes - the PRD half's ## Scope > ### Out of scope / Non-Goals (the rubric's home, SEC-006/QA-006) and the template's ## 4. Out of scope / non-goals - reconciled by a pointer line in all nine specs shipped this run. Per the recorded IMP-07 decision, the template drops section 4 and renumbers Protected invariants from 5 to 4; the rubric needs no change because it never required section 4.

Problem

Mandated duplication is authoring tax and drift risk: two homes for one statement invite divergence, and every audit this run logged the reconciliation as a wontfix finding.

Proposed Solution

Remove the ## 4. Out of scope / non-goals block from TASK-TEMPLATE.md, renumber ## 5. Protected invariants... to ## 4., and update the template schema test to assert the new shape (single out-of-scope home, invariants at 4). Existing specs are untouched - the rubric accepts both shapes since section 4 was never a rule.

Alternatives Considered

• Make section 4 optional with guidance. Rejected by the decision: an optional duplicate still invites divergence.
• Keep both and document the idiom. Rejected: nine specs of pointer lines is the evidence against it.

Success Metrics

• Primary: a spec authored from the new template carries exactly one out-of-scope home and lints clean (schema test asserts the shape on every run). Baseline: two homes plus a pointer line in 9 of 9 specs this run. Deadline: final acceptance.
• Guardrail: the payload-vendored template matches source byte-for-byte after rebuild (existing template gates keep covering it).

Scope

In scope: the template block removal, renumbering, schema-test update.

Out of scope / Non-Goals

• Rewriting existing specs (both shapes stay valid to the rubric).
• The per-type body templates (feature/bug/chore carry the PRD half only - no section 4 exists there).

Dependencies

• None; cone is the template file plus its schema test (disjoint from all batch members).

AI Authorship Disclosure

• Tools used: Claude (Fable 5) running the CyberOS task-author skill inside Cowork.
• Scope: spec drafted from the recorded IMP-07 decision; implementation under ship-tasks supervision.
• Human review: batch-3 PLAN approved 2026-07-16; both HITL gates are recorded human verdicts.

1. Description (normative)

• 1.1 TASK-TEMPLATE.md MUST NOT contain a ## 4. Out of scope / non-goals section; ## Scope > ### Out of scope / Non-Goals remains the single home.
• 1.2 The Protected-invariants section MUST be renumbered to ## 4. with its content unchanged.
• 1.3 test_template_schema.sh MUST assert the new shape: no duplicate out-of-scope heading in TASK-TEMPLATE.md, invariants present at section 4.
• 1.4 The rebuilt payload MUST carry the updated template (covered by the existing vendor gates; asserted once in the schema test against a scratch build).

2. Acceptance criteria

• [ ] AC 1 (traces_to: #1.1, #1.2) - template shape: single out-of-scope home, invariants at 4 - test: scripts/tests/test_template_schema.sh::t08_single_out_of_scope_home
• [ ] AC 2 (traces_to: #1.3) - schema test fails on a reintroduced duplicate (fixture) - test: scripts/tests/test_template_schema.sh::t08_duplicate_reintroduction_fails
• [ ] AC 3 (traces_to: #1.4) - scratch payload carries the new shape - test: scripts/tests/test_template_schema.sh::t08_payload_carries_shape

3. Edge cases

• Existing specs with the old shape: untouched and still rubric-valid; the schema test targets the TEMPLATE, never the corpus (AC 1).
• Downstream prose citing "section 5 invariants": grep at implement time and update only template-adjacent references, not historical specs.
• Security-class: none - prompt-text template edit gated by tests.