instruction-reviewlisted

<!-- intentional divergence from Supervisor: Edition expands gate names inline, replaces Supervisor-specific paths (`scripts/brain-qa.cjs`, ADR-006/007, `docs/adrs/`, `docs/ledgers/audits/`) with heir-portable language ("a brain-qa validator", "where your heir tracks framework-level decisions"). Same six gates, same criteria. Audited 2026-05-31. --> # Instruction Review Audit a candidate instruction against the five-gate contract plus optional Gate 6 (Token Budget). Instructions are always-on or pattern-applied rules — every match loads tokens, so the budget gate matters more than for any other artifact type. ## When to Use Three fire contexts: 1. **Author self-audit** — dogfood your own draft before committing. Invoked from [instruction-creator](../instruction-creator/SKILL.md) Phase 7. 2. **External candidate adoption** — gate before pulling a Mall instruction or store instruction into this brain. 3. **Periodic re-audit** — re-check existing instructions on a cadence (e.g., quarterly retraining). The shared gate model lives in [skill-review/SKILL.md § The Five Gates](../skill-review/SKILL.md) as the canonical contract. This file documents the **instruction-specific criteria** for each gate. If a gate definition here disagrees with skill-review, the gate *concept* in skill-review wins; this file owns only the per-type criteria. A mechanical validator (e.g., a brain-qa script that ships with the heir) typically checks Gate 1 frontmatter compliance and the mechanical su