arckit-cli 6.1.1__py3-none-any.whl

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/adm-preliminary.md

@@ -0,0 +1,72 @@
+# TOGAF ADM Preliminary Playbook
+
+> **Guide Origin**: Community | **ArcKit Version**: [VERSION]
+
+`/arckit:adm-preliminary` creates the Architecture Vision artefact for a TOGAF ADM engagement. Use it to set scope, drivers, constraints, success criteria, stakeholders, and ADM phase boundaries before producing business, application, gap, transition, and governance artefacts.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Global architecture principles (`ARC-000-PRIN-v1.0.md`) | Mandatory decision framework and standards baseline |
+| Stakeholder analysis (`ARC-<id>-STKE-v1.0.md`) | Drivers, goals, conflicts, outcomes, engagement needs |
+| Architecture strategy (`ARC-<id>-STRAT-v1.0.md`) | Strategic themes, investment envelope, target direction |
+| External vision or mandate documents | Existing transformation charter, EA mandate, business strategy |
+
+---
+
+## Command
+
+```bash
+/arckit:adm-preliminary <project ID or transformation scope>
+```
+
+Output: `projects/<id>/ARC-<id>-ADMP-v1.0.md`
+
+---
+
+## Deliverable Snapshot
+
+| Section | Contents |
+|---------|----------|
+| Architecture Vision | Narrative for the transformation or ADM cycle |
+| Scope | In-scope and out-of-scope capabilities, systems, and business areas |
+| Drivers | Strategic, operational, compliance, and technology drivers |
+| Constraints | Budget, timeline, regulatory, technical, and resource constraints |
+| Success Criteria | Measurable targets for the ADM engagement |
+| Architecture Landscape | High-level Mermaid context diagram |
+| Stakeholder Map | Stakeholder interest, influence, and engagement approach |
+| ADM Scope | Phase coverage and exclusions |
+| Traceability | Links from principles, stakeholders, and strategy into the vision |
+
+---
+
+## Workflow
+
+| Step | Command | Result |
+|------|---------|--------|
+| 1 | `/arckit:principles` | Establish global architecture guardrails |
+| 2 | `/arckit:stakeholders` | Capture business drivers and stakeholder goals |
+| 3 | `/arckit:strategy` | Set the strategic direction and investment frame |
+| 4 | `/arckit:adm-preliminary` | Define the ADM scope and Architecture Vision |
+| 5 | `/arckit:business-capability-map` | Translate the vision into business capabilities |
+
+---
+
+## Review Checklist
+
+- Architecture principles are present and referenced.
+- Scope boundaries are explicit enough to prevent uncontrolled expansion.
+- At least three drivers and three measurable success criteria are included.
+- ADM phases are marked in scope or out of scope with rationale.
+- External strategy or mandate sources are cited where used.
+
+---
+
+## Related Commands
+
+- `/arckit:business-capability-map` - Map capabilities within the ADM scope.
+- `/arckit:gap-analysis` - Compare current and target architecture states.
+- `/arckit:architecture-board` - Establish the governance body for implementation.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/adr.md

@@ -0,0 +1,60 @@
+# Architecture Decision Record (ADR) Guide
+
+> **Guide Origin**: Official | **ArcKit Version**: [VERSION]
+
+`/arckit:adr` produces MADR v4.0-compatible architecture decision records backed by governance metadata and UK Government escalation requirements.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Requirements & principles | Provide evaluation criteria and constraints |
+| Risk register | Ensures mitigations or accepts residual risk |
+| HLD/DLD references | Supplies context diagrams and interfaces |
+| Compliance artefacts | Link to TCoP, Secure by Design, AI Playbook evidence when impacted |
+
+---
+
+## Command
+
+```bash
+/arckit:adr Document architectural decision for <topic>, compare options A/B/C
+```
+
+Output: `projects/<id>/decisions/ARC-<id>-ADR-001-v1.0.md` (uses multi-instance numbering) plus optional summary table appended to `decisions/index.md`.
+
+> **Auto-versioning**: Creating a new ADR auto-increments the ADR number. Updating an existing ADR (e.g., "update ADR-001") auto-increments the version (minor for updated evidence, major for changed decision outcome) instead of overwriting.
+
+---
+
+## MADR Structure
+
+| Section | Notes |
+|---------|-------|
+| Context & Problem Statement | State trigger, affected capabilities, and non-negotiables |
+| Decision Drivers | Reference stakeholder goals, NFRs, and compliance mandates |
+| Considered Options | Minimum of two options; include baseline / do-nothing |
+| Pros & Cons | Map to decision drivers; cite evidence from research or pilots |
+| Decision Outcome | Selected option, rationale, status (Proposed/Accepted/Rejected) |
+| Consequences | Positive/negative impacts, follow-on tasks, re-evaluation date |
+| Governance & Approvals | Record design authorities, CABs, risk owners, sign-off dates |
+
+---
+
+## Usage Pattern
+
+1. Run `/arckit:adr` immediately after new insights from `/arckit:research`, `/arckit:hld-review`, or vendor clarifications.
+2. Attach document IDs from `/arckit:traceability` so requirements coverage stays auditable.
+3. Version ADRs in git; never overwrite history—open new ADRs when reversing prior choices.
+4. Feed summaries into `/arckit:story` so programme leadership tracks pivotal decisions.
+
+---
+
+## Review Checklist
+
+- Decision ties directly to stakeholder drivers or requirements IDs.
+- Risks and mitigations reference the risk register entry numbers.
+- Compliance impacts call out required evidence (TCoP point, Secure by Design activity, ATRS field).
+- Follow-on actions list named owners and due dates; backlog items reference `/arckit:backlog`.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/agent-design.md

@@ -0,0 +1,71 @@
+# AI Agent Design Playbook
+
+> **Guide Origin**: Community | **ArcKit Version**: [VERSION]
+
+`/arckit:agent-design` creates an Agent Architecture Specification. It documents agent pattern, tool contracts, memory architecture, orchestration, guardrails, testing strategy, and traceability back to requirements, risks, stakeholders, and inventory.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Agent inventory (`ARC-<id>-AAGI-v1.0.md`) | Existing agents, capability baseline, ownership, operational context |
+| Requirements (`ARC-<id>-REQ-v1.0.md`) | Functional and non-functional requirements for agent capabilities |
+| Global principles (`ARC-000-PRIN-v1.0.md`) | AI, technology, and governance constraints |
+| Stakeholder analysis (`ARC-<id>-STKE-v1.0.md`) | Agent users, operators, approvers, and affected groups |
+| Risk register (`ARC-<id>-RISK-v1.0.md`) | AI, tool, security, reliability, and compliance risks |
+| AI Playbook assessment (`ARC-<id>-AIPB-v1.0.md`) | Ethics, fairness, transparency, and model governance evidence |
+
+---
+
+## Command
+
+```bash
+/arckit:agent-design <agent name or project ID>
+```
+
+Output: `projects/<id>/ARC-<id>-AAGR-v1.0.md`
+
+---
+
+## Design Patterns
+
+| Pattern | Use When |
+|---------|----------|
+| Single Agent | Focused domain, small tool set, simple operational model |
+| Chain | Sequential specialist steps with clear handoff boundaries |
+| Multi-Agent | Parallel workers or domain agents need coordination |
+| Hierarchical | Supervisor and workers need dynamic task allocation |
+
+---
+
+## Deliverable Snapshot
+
+| Section | Contents |
+|---------|----------|
+| Agent Architecture Overview | Pattern, components, capabilities, C4-style diagram |
+| Tool Contracts | Tool ID, permission, input schema, output schema, validation |
+| Memory Architecture | Session, durable, and vector memory design |
+| Orchestration Design | Routing, coordination, retries, and escalation |
+| Guardrail Configuration | Input, context, tool, output, and policy controls |
+| Testing Strategy | Unit, integration, security, regression, and acceptance tests |
+| Traceability | Links to inventory, requirements, risks, stakeholders, and principles |
+
+---
+
+## Review Checklist
+
+- Agent purpose and pattern are explicit and justified.
+- Tool contracts include permissions, schemas, and failure behaviour.
+- Memory design separates session, durable, and vector responsibilities.
+- Guardrails are mapped to real requirements and risks.
+- Testing covers tool contracts, orchestration, prompt injection, and output validation.
+
+---
+
+## Related Commands
+
+- `/arckit:agent-integration` - Design multi-agent contracts and shared state.
+- `/arckit:agent-security` - Add sandboxing, permission, and validation controls.
+- `/arckit:agent-governance` - Define oversight and approvals.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/agent-governance.md

@@ -0,0 +1,61 @@
+# AI Agent Governance Playbook
+
+> **Guide Origin**: Community | **ArcKit Version**: [VERSION]
+
+`/arckit:agent-governance` creates an Agent Governance Framework covering oversight model, approval workflow, audit requirements, monitoring KPIs, escalation procedures, incident response, and compliance mapping.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Agent inventory (`ARC-<id>-AAGI-v1.0.md`) | Mandatory agent list, capabilities, risk levels |
+| Agent design (`ARC-<id>-AAGR-v1.0.md`) | Mandatory autonomy, tool use, decision patterns, integration points |
+| Global principles (`ARC-000-PRIN-v1.0.md`) | Governance principles and human oversight standards |
+| Architecture board (`ARC-<id>-BORD-v1.0.md`) | Board policies, risk appetite, approval precedents |
+| Risk register (`ARC-<id>-RISK-v1.0.md`) | AI, operational, security, and compliance risks |
+| AI or regulatory assessments | AI Playbook, AI Act, NIST AI RMF, sector controls as applicable |
+
+---
+
+## Command
+
+```bash
+/arckit:agent-governance <agent, project ID, or agent programme>
+```
+
+Output: `projects/<id>/ARC-<id>-AAOV-v1.0.md`
+
+---
+
+## Deliverable Snapshot
+
+| Section | Contents |
+|---------|----------|
+| Oversight Model | Human oversight tiers by risk and autonomy |
+| Approval Matrix | Who approves design, deployment, tool access, and changes |
+| Audit Requirements | Logs, evidence, retention, review cadence |
+| Monitoring KPIs | Safety, quality, performance, incident, and drift signals |
+| Escalation Procedures | Operational and governance escalation paths |
+| Incident Response Plan | Severity, response times, owners, communications |
+| Compliance Mapping | Applicable AI, security, data, and sector controls |
+| Traceability | Links to inventory, design, principles, board, and risk artefacts |
+
+---
+
+## Review Checklist
+
+- Oversight model matches autonomy, risk, and user impact.
+- Approval matrix covers deployment, tool access, model changes, and policy exceptions.
+- Audit requirements include retention, reviewers, and evidence location.
+- Monitoring KPIs are actionable and assigned to owners.
+- Compliance mapping distinguishes required controls from advisory practices.
+
+---
+
+## Related Commands
+
+- `/arckit:agent-security` - Align security controls to governance requirements.
+- `/arckit:architecture-board` - Route high-impact agent decisions.
+- `/arckit:agent-maturity` - Assess programme maturity and improvement roadmap.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/agent-integration.md

@@ -0,0 +1,59 @@
+# AI Agent Integration Playbook
+
+> **Guide Origin**: Community | **ArcKit Version**: [VERSION]
+
+`/arckit:agent-integration` creates an Agent Integration Architecture for multi-agent systems. It defines integration scope, inter-agent contracts, message protocol, shared state, failure isolation, observability, and traceability.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Agent design (`ARC-<id>-AAGR-v1.0.md`) | Mandatory agent definitions and interface requirements |
+| Agent inventory (`ARC-<id>-AAGI-v1.0.md`) | Agent register, lifecycle, and deployment context |
+| Agent governance (`ARC-<id>-AAOV-v1.0.md`) | Oversight, approval, and audit requirements |
+| Agent security (`ARC-<id>-AASE-v1.0.md`) | Security controls and permitted communication patterns |
+| System diagrams or HLD reviews | Existing system integration and platform context |
+
+---
+
+## Command
+
+```bash
+/arckit:agent-integration <project ID or agent group>
+```
+
+Output: `projects/<id>/ARC-<id>-AAIN-v1.0.md`
+
+---
+
+## Deliverable Snapshot
+
+| Section | Contents |
+|---------|----------|
+| Integration Architecture | Agent topology, orchestration, and boundaries |
+| Inter-Agent Contracts | Contract name, sender, receiver, schema, SLA, failure mode |
+| Message Protocol | Transport, envelope, correlation, idempotency, retries |
+| Shared State Design | State ownership, locking, retention, consistency model |
+| Failure Isolation | Circuit breakers, fallback, dead-letter paths, containment |
+| Observability | Logs, traces, metrics, alerts, audit linkage |
+| Traceability | Links to designs, inventory, governance, and security controls |
+
+---
+
+## Review Checklist
+
+- Every integration has a named contract and accountable owner.
+- Message schemas include versioning and compatibility expectations.
+- Shared state ownership is explicit and avoids hidden coupling.
+- Failure isolation contains bad outputs, tool failures, and cross-agent injection.
+- Observability supports operational debugging and governance audit.
+
+---
+
+## Related Commands
+
+- `/arckit:agent-design` - Define individual agent architecture and interfaces.
+- `/arckit:agent-governance` - Apply oversight to multi-agent orchestration.
+- `/arckit:agent-security` - Secure inter-agent communication.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/agent-inventory.md

@@ -0,0 +1,60 @@
+# AI Agent Inventory Playbook
+
+> **Guide Origin**: Community | **ArcKit Version**: [VERSION]
+
+`/arckit:agent-inventory` creates a governed catalogue of AI agents across a project, programme, or organisation. It records agent purpose, capabilities, ownership, lifecycle, dependencies, security classification, and human oversight level.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| ADM preliminary (`ARC-<id>-ADMP-v1.0.md`) | Optional programme scope and transformation context |
+| Application inventory (`ARC-<id>-APP-v1.0.md`) | Application and hosting context for deployed agents |
+| Global principles (`ARC-000-PRIN-v1.0.md`) | Agent governance, AI, technology, and risk guardrails |
+| Existing agent designs (`ARC-<id>-AAGR-v1.0.md`) | Agent capabilities and architecture details |
+| Agent governance (`ARC-<id>-AAOV-v1.0.md`) | Oversight tiers and approval model |
+| Agent security (`ARC-<id>-AASE-v1.0.md`) | Security posture and controls |
+
+---
+
+## Command
+
+```bash
+/arckit:agent-inventory <project ID, organisation scope, or environment>
+```
+
+Output: `projects/<id>/ARC-<id>-AAGI-v1.0.md`
+
+---
+
+## Deliverable Snapshot
+
+| Section | Contents |
+|---------|----------|
+| Agent Register | Name, owner, purpose, lifecycle, environment |
+| Capability Matrix | Business and technical capabilities per agent |
+| Agent Dependencies | Models, tools, data sources, systems, other agents |
+| Security Classification | Data sensitivity, tool risk, access, deployment exposure |
+| Agent Lifecycle | Prototype, pilot, production, deprecated, retired |
+| Oversight Model | Human-in-the-loop, human-on-the-loop, escalation path |
+| Traceability | Links to ADM, applications, design, governance, and security artefacts |
+
+---
+
+## Review Checklist
+
+- Inventory scope is explicit: project, programme, environment, or enterprise.
+- Each agent has an owner and lifecycle status.
+- Tool, data, model, and system dependencies are captured.
+- Oversight level matches autonomy and risk.
+- Security classification is evidence-backed, not inferred from the agent name.
+
+---
+
+## Related Commands
+
+- `/arckit:agent-design` - Design new or modified agents.
+- `/arckit:agent-governance` - Define oversight and approval workflows.
+- `/arckit:agent-security` - Assess and design agent security controls.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/agent-maturity.md

@@ -0,0 +1,84 @@
+# AI Agent Maturity Playbook
+
+> **Guide Origin**: Community | **ArcKit Version**: [VERSION]
+
+`/arckit:agent-maturity` assesses an AI agent programme across design, governance, security, integration, and operations. It produces a current-state maturity view, target maturity, improvement roadmap, benchmarks, and traceability.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Agent inventory (`ARC-<id>-AAGI-v1.0.md`) | Agent baseline, ownership, lifecycle, capability context |
+| Agent design (`ARC-<id>-AAGR-v1.0.md`) | Design patterns, tool contracts, guardrails, orchestration |
+| Agent governance (`ARC-<id>-AAOV-v1.0.md`) | Oversight, audit, monitoring, approval process |
+| Agent security (`ARC-<id>-AASE-v1.0.md`) | Security controls, threat model, validation pipeline |
+| Agent integration (`ARC-<id>-AAIN-v1.0.md`) | Integration patterns and interoperability evidence |
+| Risk register (`ARC-<id>-RISK-v1.0.md`) | Operational, compliance, and AI risk context |
+| Enterprise maturity model | Wider maturity baseline and benchmark context |
+
+---
+
+## Command
+
+```bash
+/arckit:agent-maturity <project ID or agent programme>
+```
+
+Output: `projects/<id>/ARC-<id>-AAMT-v1.0.md`
+
+---
+
+## Maturity Dimensions
+
+| Dimension | Evidence |
+|-----------|----------|
+| Design | Reusable patterns, tool contracts, design reviews, ADRs |
+| Governance | Oversight model, audit coverage, approval records, compliance mapping |
+| Security | Threat model, sandboxing, permission controls, security testing |
+| Integration | Standard APIs, contracts, observability, failure isolation |
+| Operations | Monitoring, incident response, SLAs, support model, continuous improvement |
+
+---
+
+## Maturity Levels
+
+| Level | Meaning |
+|-------|---------|
+| L1 Ad-hoc | Reactive, informal, undocumented |
+| L2 Reactive | Some process after issues occur |
+| L3 Defined | Standard process documented and repeatable |
+| L4 Managed | Metrics-driven and actively governed |
+| L5 Optimized | Predictive, continuously improved, evidence-led |
+
+---
+
+## Deliverable Snapshot
+
+| Section | Contents |
+|---------|----------|
+| Maturity Model Framework | Dimensions, levels, and evidence criteria |
+| Current State Assessment | Level, evidence, gaps, and risk per dimension |
+| Target State | Target level and rationale per dimension |
+| Improvement Roadmap | Initiatives, owners, timeline, dependencies, success criteria |
+| Benchmarks | Industry, regulatory, and peer comparisons |
+| Traceability | Links to inventory, design, governance, security, integration, and roadmap artefacts |
+
+---
+
+## Review Checklist
+
+- Ratings cite evidence rather than relying on opinion.
+- Target levels are realistic for the investment and timeline.
+- Each improvement initiative has owner, success criteria, and dependency notes.
+- Security and governance maturity are not averaged away by strong design maturity.
+- Roadmap actions feed back into `/arckit:agent-design`, `/arckit:agent-governance`, and `/arckit:agent-security`.
+
+---
+
+## Related Commands
+
+- `/arckit:maturity-model` - Assess wider enterprise architecture maturity.
+- `/arckit:agent-design` - Address design maturity gaps.
+- `/arckit:agent-governance` - Address governance maturity gaps.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/agent-security.md

@@ -0,0 +1,63 @@
+# AI Agent Security Playbook
+
+> **Guide Origin**: Community | **ArcKit Version**: [VERSION]
+
+`/arckit:agent-security` creates an Agent Security Architecture covering threat model, sandboxing, tool permissions, data handling, prompt-injection defences, output validation, secret management, incident response, and traceability.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Agent inventory (`ARC-<id>-AAGI-v1.0.md`) | Mandatory agent list, capabilities, risk levels |
+| Agent design (`ARC-<id>-AAGR-v1.0.md`) | Mandatory tool contracts, memory, guardrails, integration points |
+| Secure by Design (`ARC-<id>-SECD-v1.0.md`) | Existing security controls and threat model |
+| Risk register (`ARC-<id>-RISK-v1.0.md`) | Security, AI, tool, and data risks |
+| Global principles (`ARC-000-PRIN-v1.0.md`) | Security and technology constraints |
+| Agent governance (`ARC-<id>-AAOV-v1.0.md`) | Oversight tiers and compliance requirements |
+| DPIA or AI Playbook artefacts | Personal data and AI ethics implications |
+
+---
+
+## Command
+
+```bash
+/arckit:agent-security <agent or project ID>
+```
+
+Output: `projects/<id>/ARC-<id>-AASE-v1.0.md`
+
+---
+
+## Deliverable Snapshot
+
+| Section | Contents |
+|---------|----------|
+| Threat Model | Input, execution, output, and infrastructure attack surfaces |
+| Sandboxing Architecture | Agent, tool, memory, network, and data isolation |
+| Tool Permission Matrix | Permissions, access scope, risk, and mitigation by tool |
+| Data Handling Policy | PII, sensitive, public, model, and secret handling rules |
+| Prompt Injection Defences | Input, context, output, and tool-call controls |
+| Output Validation Pipeline | Schema, safety, PII, human review, release stages |
+| Secret Management | Storage, scope, rotation, and ownership |
+| Incident Response | Severity tiers, owners, actions, and timing |
+| Traceability | Links to inventory, design, governance, and Secure by Design artefacts |
+
+---
+
+## Review Checklist
+
+- Threat model covers direct and indirect prompt injection.
+- High-risk tools have explicit permission checks and mitigations.
+- Sandboxing boundaries are enforceable, not just policy statements.
+- Output validation includes schema, content, and PII checks.
+- Secret lifecycle includes storage, scope, owner, and rotation cadence.
+
+---
+
+## Related Commands
+
+- `/arckit:agent-governance` - Align security controls with oversight model.
+- `/arckit:secure` - Produce wider Secure by Design evidence.
+- `/arckit:agent-integration` - Secure communication between agents.

arckit_cli-6.1.1.data/data/share/arckit/docs/guides/ai-playbook.md

@@ -0,0 +1,103 @@
+# AI Playbook Assessment Playbook
+
+> **Guide Origin**: Official | **ArcKit Version**: [VERSION]
+
+`/arckit:ai-playbook` assesses UK Government AI Playbook compliance for responsible AI deployment in public sector projects.
+
+---
+
+## Inputs
+
+| Artefact | Purpose |
+|----------|---------|
+| Requirements (`ARC-<id>-REQ-v1.0.md`) | AI-related functional and non-functional requirements |
+| Architecture diagrams | AI system components and data flows |
+| Data model | Training data sources, personal data processing |
+| Stakeholder drivers | Business context and user impact |
+
+---
+
+## Command
+
+```bash
+/arckit:ai-playbook Assess AI Playbook compliance for <project>
+```
+
+Output: `projects/<id>/ARC-<id>-AIPB-v1.0.md`
+
+---
+
+## Assessment Structure
+
+| Section | Contents |
+|---------|----------|
+| Executive Summary | Overall compliance status and key findings |
+| AI System Overview | System description, use cases, decision types |
+| 10 Principles Assessment | Compliance against each AI Playbook principle |
+| Risk Assessment | AI-specific risks (bias, transparency, accountability) |
+| Data Governance | Training data quality, bias mitigation, personal data |
+| Human Oversight | Human-in-the-loop requirements, escalation paths |
+| Explainability | How decisions can be explained to users |
+| Monitoring & Evaluation | Ongoing performance and fairness monitoring |
+| Compliance Roadmap | Actions needed for full compliance |
+
+---
+
+## 10 AI Playbook Principles
+
+| # | Principle | Focus Area |
+|---|-----------|------------|
+| 1 | Use AI to support public sector tasks | Legitimate public benefit |
+| 2 | Have clear goals and robust evaluation | Measurable outcomes |
+| 3 | Be fair, accountable and transparent | Bias mitigation, explainability |
+| 4 | Consider the full lifecycle | Development through retirement |
+| 5 | Use quality data responsibly | Data governance and ethics |
+| 6 | Ensure adequate human oversight | Human-in-the-loop controls |
+| 7 | Keep data secure and manage risks | Security and risk management |
+| 8 | Use existing tools where appropriate | Avoid reinvention |
+| 9 | Build internal capability | Skills and knowledge transfer |
+| 10 | Work in the open | Transparency and collaboration |
+
+---
+
+## One-Page Workflow
+
+| Phase | Key Activities | ArcKit Commands |
+|-------|----------------|-----------------|
+| Discovery | Define AI use cases and requirements | `/arckit:requirements`, `/arckit:stakeholders` |
+| Data | Design data model and assess data quality | `/arckit:data-model`, `/arckit:dpia` |
+| Governance | Assess AI Playbook compliance | `/arckit:ai-playbook` |
+| Transparency | Create algorithmic transparency record | `/arckit:atrs` |
+| Operations | Plan ongoing monitoring and governance | `/arckit:mlops`, `/arckit:operationalize` |
+
+---
+
+## Review Checklist
+
+- All 10 AI Playbook principles assessed with evidence.
+- Bias risks identified with mitigation measures.
+- Human oversight requirements clearly defined.
+- Explainability approach documented for affected users.
+- Training data quality and governance addressed.
+- Ongoing monitoring plan for fairness and performance.
+- Compliance roadmap with owners and timelines.
+
+---
+
+## Related UK Government Requirements
+
+| Requirement | Document | When Needed |
+|-------------|----------|-------------|
+| Algorithmic Transparency | `ARC-<id>-ATRS-v1.0.md` | Public-facing algorithmic decisions |
+| Data Protection Impact | `ARC-<id>-DPIA-v1.0.md` | AI processing personal data |
+| MOD AI Assurance | `ARC-<id>-JSP936-v1.0.md` | Defence AI systems |
+
+---
+
+## Key Principles
+
+1. **Public Benefit First**: AI must serve legitimate public sector goals.
+2. **Fairness by Design**: Bias detection and mitigation from the start.
+3. **Transparency**: Users should understand how AI affects them.
+4. **Human Oversight**: Appropriate human control based on risk level.
+5. **Continuous Evaluation**: Ongoing monitoring, not just initial assessment.