Architecture

Quick Summary

Review modes: For review/audit work, standalone defaults to --review-only (--report-only alias); --fix-loop enables review → validate → authorized fix → fresh re-review, default cap 3. --fix-loop --loop-owner=caller returns a read-only pass to the caller that owns fixes/re-review. Create triage, review, validation and re-review todo tasks first; apply the carried review-policy contract for LOW deferral and user-approved bounded extensions. Non-review modes and terminal findings validation keep their own dispatch.

Goal: Run the selected architecture contract to design a solution, review compliance, grade scalability, or deliver a whole-project health report with evidence-backed decisions.

Summary: Select the explicit --mode → read its complete reference → execute its own phases, evidence gates, flags and outputs → validate and hand off. No mode prints the table and stops, without questions or work. full composes read-only reviewer faces; other modes run independently.

Workflow: Dispatch → load the selected contract → triage targets and plan review tasks → execute the mode → validate → hand off.

Key Rules:

• Detect --mode before any work; load only the selected reference in full.
• Preserve each mode's scope, authority, scores, gates, report paths and round caps.
• For capacity/cache decisions, apply the selected mode's measured-capacity protocol: workload evidence and business availability needs govern the next step.

Mode Dispatch

Read the named reference in full when its explicit mode is selected. No mode: show this table and stop; ask nothing, run nothing, never infer a default.

• init / audit (also mode=init / mode=audit) are scalability run types, separate from --mode=scalability. Read references/scorecard.md when scoring that mode.
• Only full composes modes: its children read and run the scalability and review references plus the separate production-readiness-review skill. Never copy or re-implement a face.

<!-- SYNC:review-decision-autonomy:reminder -->

MUST ATTENTION Decide supported review choices and recommendations, record the rationale, and finish without routine user questions. Keep round-limit extension, indispensable facts and actual action authority; preserve source coverage, validation, tests, read-only boundaries and budgets. Review decisions do not accept open risks or make a failed gate pass.

<!-- /SYNC:review-decision-autonomy:reminder -->

Closing Reminders

IMPORTANT MUST ATTENTION Goal: Run the selected architecture contract to design a solution, review compliance, grade scalability, or deliver a whole-project health report with evidence-backed decisions.

MUST ATTENTION Main steps: dispatch explicit mode → read its complete contract → triage targets and plan review tasks → execute the mode's phases → validate → hand off.

• No mode prints the table and stops: no question, inferred mode or audit.
• Load only the selected reference; retain its flags, gates, report paths, score semantics, authority and round caps.
• full runs INLINE and composes read-only sub-agent faces; each face loads its own contract.
• Capacity decisions follow workload → safe measured-work reduction → capacity/recovery proof, within the selected mode's contract.

Anti-Rationalization:

<!-- PROTOCOL-GUIDES:START -->

Protocol guides — Apply each protocol only when its read-when condition holds. If a protocol's text is not in your context, read its file below before you act on it.

• review-decision-autonomy — Choose supported review decisions and ask before extending the round budget; running any review or audit skill or mode → .claude/skills/shared/protocols/review-decision-autonomy.md
• review-policy — Review-only or a three-round fix loop with explicit extension and fresh post-fix evidence; deciding round eligibility, blocking findings or review state → .claude/skills/shared/protocols/review-policy.md

<!-- PROTOCOL-GUIDES:END -->

<!-- SYNC:review-policy:reminder -->

MUST ATTENTION Triage all targets, plan and create review/validation/fix/fresh re-review tasks first. Review-only reports without edits; fix-loop freshly reviews every repair under one fixing owner. Default cap three rounds: disclose deferred LOWs, ask and wait before a bounded extension for MEDIUM+ or failed required checks. Preserve scope, coverage and actual operation authority.

<!-- /SYNC:review-policy:reminder -->