Advanced Governance-Automation Criteria

Compatibility note: this file keeps the historical docs/v8-scope.md path and V8-linked internal identifiers for scripts, tools, and tests. Public product language should describe this as an advanced governance-automation maturity layer, not as a product version.

This criteria layer remains local-first. It excludes hosted service work, database/storage work, GitHub/PR adapters, CLI UX, and typed SDKs.

It should make architect-mcp better at local governance automation, standards refactoring, and resilient agent collaboration while staying inside the existing MCP/domain/test architecture.

Governance-Automation Direction

This layer focuses on local governance automation:

• Standards refactoring.
• Policy minimization.
• Local review playbooks.
• Agent collaboration protocols.
• Failure-mode drills.
• Rule impact calibration.
• Documentation intelligence.

Standards Refactoring

Goal: help maintainers improve existing packs and policies without changing runtime architecture.

Deliverables:

• Detect duplicate, overlapping, stale, or weak rules across packs.
• Suggest rule splits, merges, renames, and severity changes.
• Identify rules with weak examples or vague remediation.
• Suggest detector families for manual-only rules.
• Generate before/after pack diffs with risk notes.

Policy Minimization

Goal: help projects avoid overloading agents with too many rules.

Deliverables:

• Minimal policy set recommendations for a brief and file summary.
• Rule usefulness scoring based on triggered signals and selected stack.
• Noise forecast when too many packs apply.
• "Drop, defer, or keep" recommendations for candidate rules.
• Compact standards profile output for token-sensitive clients.

Local Review Playbooks

Goal: turn recurring review situations into repeatable local workflows.

Deliverables:

• Playbooks for fresh app generation, feature addition, refactor, security-sensitive change, dependency update, UI polish, and test hardening.
• Each playbook defines intake checks, pre-edit contract fields, relevant policy packs, review tools, and proof requirements.
• Playbook selection from request text and project brief.
• Playbook conformance review after implementation.

Agent Collaboration Protocols

Goal: make multi-agent or multi-step local work safer without subagent orchestration, hosted state, or external systems.

Deliverables:

• Handoff protocol templates.
• File ownership declaration checks.
• Parallel-work conflict warnings from proposed file plans.
• "Do not touch" boundary declarations.
• Integration checklist for merging independently produced local changes.

Failure-Mode Drills

Goal: test whether the harness catches known bad agent behavior before release.

Deliverables:

• Drill fixtures for skipped verification, fake root-cause claims, dependency churn, broad rewrites, misplaced secrets, UI-to-server leaks, and rule overreach.
• Drill runner output that explains what was caught and what escaped.
• Suggested new detectors or evals for escaped failures.

Rule Impact Calibration

Goal: tune severity, confidence, and gate impact locally.

Deliverables:

• Compare severity settings across policy profiles.
• Identify rules that are too noisy or too weak for a project archetype.
• Recommend confidence adjustments from available evidence.
• Preview gate result changes from severity/confidence edits.
• Keep calibration stateless and based only on supplied reports or fixtures.

Documentation Intelligence

Goal: keep human and agent docs aligned with actual tool behavior.

Deliverables:

• Detect stale claims in README, llms.txt, AGENTS.md, and generated docs.
• Summarize tool surface changes into doc update suggestions.
• Identify missing examples for new tools or rules.
• Validate that roadmap docs match tested scope boundaries.
• Generate concise release notes from supplied local change summaries.

Non-Goals

• Hosted service or hosted API implementation.
• Database/storage, including schema, persistence, migrations, review history, or memory storage.
• GitHub repository, branch, PR, issue, app, OAuth, or comment workflow.
• CLI UX.
• Typed client SDK.
• Billing, accounts, teams, dashboards, or remote policy management.

Those remain out of scope until a later backend/productization phase.