ArcKit v0.2.2 - OpenAI Codex CLI Support & Enhanced Stakeholder Analysis

Release Date: 2025-10-20

🎉 Major Features

🤖 OpenAI Codex CLI Support

ArcKit now supports OpenAI Codex CLI in addition to Claude Code! Codex CLI is OpenAI's coding agent included with ChatGPT Plus, Pro, Business, Edu, and Enterprise plans.

Why This Matters:

✅ Same powerful workflows on OpenAI's platform
✅ No additional infrastructure - Codex CLI already exists
✅ Included with ChatGPT Plus ($20/month, no API costs)
✅ Local file system access with bash script execution
✅ Custom slash commands via .codex/prompts/

How to Use:

# Set CODEX_HOME for project-specific commands
export CODEX_HOME="$(pwd)/.codex"
codex --auto

# Use ArcKit commands with /prompts: prefix
/prompts:arckit.principles Create principles for financial services
/prompts:arckit.stakeholders Analyze stakeholders for cloud migration
/prompts:arckit.requirements Create comprehensive requirements

What's Included:

📁 .codex/ folder structure with 17 prompts
📚 Comprehensive .codex/README.md setup guide (400+ lines)
🔧 Approval modes documentation (--auto, --read-only, --network)
🆚 Comparison guide: Codex CLI vs Claude Code
🐛 Troubleshooting section

Supported AI Agents (now 5 total):

Claude Code ✅ (Recommended)
OpenAI Codex CLI ✅ (NEW in v0.2.2)
GitHub Copilot ✅
Cursor ✅
Gemini CLI ✅

👥 Enhanced Stakeholder Analysis (v0.2.1)

Comprehensive stakeholder analysis framework that identifies who cares about your project and why, with complete traceability from drivers to outcomes.

New Command: /arckit.stakeholders

/arckit.stakeholders Analyze stakeholders for cloud migration where CFO wants cost savings, CTO wants innovation, and Operations is worried about downtime

Key Features:

1. Power-Interest Grid

Classify stakeholders by engagement strategy:

Manage Closely: High power, high interest (key players)
Keep Satisfied: High power, low interest (influential)
Keep Informed: Low power, high interest (advocates)
Monitor: Low power, low interest (minimal engagement)

2. Seven Types of Drivers

Deep analysis of what motivates each stakeholder:

STRATEGIC: Competitive advantage, market position, innovation
OPERATIONAL: Efficiency, quality, speed, reliability
FINANCIAL: Cost reduction, revenue growth, ROI
COMPLIANCE: Regulatory requirements, audit findings, risk mitigation
PERSONAL: Career advancement, workload reduction, reputation
RISK: Avoiding penalties, preventing failures
CUSTOMER: Satisfaction, retention, acquisition

3. Complete Traceability

Map from motivations to measurable outcomes:

Stakeholder → Driver → Goal → Outcome
   CFO    → Cost reduction → 40% savings by Q4 → £2M annual savings

4. Conflict Analysis & Resolution

Systematic approach to handling competing stakeholder needs:

Common Conflicts:

Speed vs Quality: CFO wants fast delivery vs Operations wants thorough testing
Cost vs Features: Finance wants minimal spend vs Product wants rich features
Security vs Usability: Security wants MFA vs Users want seamless experience
Flexibility vs Standardization: Business wants customization vs IT wants standards

Resolution Strategies:

PRIORITIZE: Choose one stakeholder over another (with justification)
COMPROMISE: Find middle ground that partially satisfies both
PHASE: Satisfy both but at different times (MVP vs Phase 2)
INNOVATE: Find creative solution that satisfies both

5. RACI Matrix

Document decision authority for governance:

Responsible: Who does the work
Accountable: Who makes the final decision
Consulted: Who provides input
Informed: Who needs to know

Template: templates/stakeholder-drivers-template.md (400+ lines)

🔄 Updated Requirements Workflow

CRITICAL CHANGE: Stakeholder analysis now MUST come BEFORE requirements.

Old Workflow ❌:

Principles → Requirements → Design

New Workflow ✅:

Principles → Stakeholders → Requirements → Design

Why This Matters:

Requirements should address stakeholder goals, not invented needs
Prioritization should reflect stakeholder power/interest, not gut feel
Conflicts between requirements often stem from stakeholder conflicts

Enhanced /arckit.requirements Command:

Checks for stakeholder analysis first:

If stakeholder analysis doesn't exist:
→ Strongly recommend running /arckit.stakeholders first

Traces requirements back to stakeholder goals:

BR-001: Reduce infrastructure costs 40% by Q4 2025
Rationale: Addresses CFO's goal G-1 (cost reduction driver)
Stakeholder: CFO (High Power, High Interest)

Identifies and resolves requirement conflicts:

## Requirement Conflicts & Resolutions

### Conflict C-1: Timeline vs Quality

**Conflicting Requirements**:
- FR-001: Deliver MVP in 3 months (CEO driver: market pressure)
- NFR-Q-002: 95% test coverage (CTO driver: quality standards)

**Trade-off Analysis**:
| Option | Pros | Cons | Impact |
|--------|------|------|--------|
| Prioritize Speed | ✅ Meet market window | ❌ Higher bugs | CEO happy, CTO concerned |
| Prioritize Quality | ✅ Quality reputation | ❌ Miss market | CTO happy, CEO frustrated |
| Compromise | ✅ Balance | ❌ Neither fully satisfied | Both somewhat satisfied |
| Innovate | ✅ Automated testing | ❌ Higher cost | Both satisfied if works |

**Decision**: INNOVATE - Invest in automated testing pipeline
**Rationale**: Pays off long-term, satisfies both stakeholders
**Decision Authority**: CTO (per RACI matrix)
**Impact**: FR-001 timeline extended to 4 months, NFR-Q-002 achieved via automation

Documents stakeholder management:
- Who "won" and who "lost" in conflicts
- How to communicate with losing stakeholders
- What was deferred to future phases

Template: Updated templates/requirements-template.md with conflict resolution section

📊 Statistics

20 files changed
6,342+ lines added
1 new command (/arckit.stakeholders)
1 new folder structure (.codex/ for OpenAI Codex CLI)
17 prompts for Codex CLI users
1 comprehensive setup guide (.codex/README.md)
1 integration strategy document (OPENAI-INTEGRATION-PLAN.md)
Enhanced workflow: Stakeholders now inform requirements

🎯 Use Cases

Cloud Migration with Competing Stakeholders

# 1. Initialize project
arckit init cloud-migration --ai codex  # NEW: Codex CLI support!
cd cloud-migration

export CODEX_HOME="$(pwd)/.codex"
codex --auto

# 2. Establish principles
/prompts:arckit.principles Create cloud-first principles with FinOps

# 3. Analyze stakeholders (NEW STEP - DO THIS BEFORE REQUIREMENTS!)
/prompts:arckit.stakeholders Analyze stakeholders where:
- CFO wants £2M annual cost savings
- CTO wants modern tech stack to attract talent
- Operations Director is worried about downtime during migration
- CISO needs enhanced security controls for compliance

# Output includes:
# - Power-Interest Grid classifying all stakeholders
# - Driver analysis for each (FINANCIAL, STRATEGIC, RISK, COMPLIANCE)
# - SMART goals mapped to each driver
# - Measurable outcomes with KPIs
# - Conflict analysis: CFO speed vs Operations safety
# - Resolution strategy: Phased approach (low-risk apps first)
# - RACI matrix for decision authority

# 4. Create requirements informed by stakeholder goals
/prompts:arckit.requirements Create cloud migration requirements

# Output includes:
# - BR-001: "Reduce infrastructure costs 40% by Q4 2025"
#   Addresses CFO's goal G-1 (cost reduction driver)
# - NFR-P-001: "Maintain 99.95% uptime during migration"
#   Addresses Operations Director's goal G-3 (risk mitigation driver)
# - Requirement Conflicts section documenting:
#   * Conflict C-1: Speed (CFO) vs Safety (Operations)
#   * Resolution: PHASE - Start with low-risk apps
#   * Decision Authority: CTO (per RACI matrix)

# 5. Continue with design reviews
/prompts:arckit.hld-review Review high-level design for cloud architecture

UK Government AI Project with Complex Stakeholder Landscape

# Initialize for government AI project
arckit init benefits-chatbot --ai claude

# Analyze government stakeholders
/arckit.stakeholders Analyze stakeholders for DWP benefits chatbot where:
- Minister wants quick delivery for manifesto commitment
- Permanent Secretary needs proper governance to avoid NAO criticism
- Service Delivery wants to reduce call center volume
- Citizens need accurate, accessible answers
- ICO requires data protection compliance
- Treasury demands value for money

# Output includes UK-specific drivers:
# - POLITICAL: Minister responds to parliamentary questions
# - RISK/ACCOUNTABILITY: Permanent Secretary avoids audit findings
# - OPERATIONAL: Service Delivery improves citizen experience
# - USER: Citizens get fast, accurate service
# - REGULATORY: ICO enforces GDPR compliance
# - FINANCIAL: Treasury controls spending

# Requirements will trace back to these drivers:
# - BR-001: "Launch MVP in 6 months" (Minister's manifesto commitment)
# - NFR-C-001: "GDPR compliance with DPIA" (ICO regulatory requirement)
# - FR-001: "Answer 90% of benefits queries" (Service Delivery operational driver)
# - Conflict C-1: Speed (Minister) vs Due Diligence (Permanent Secretary)
#   Resolution: PHASE - Alpha with limited scope, Beta adds full compliance

# Create requirements aligned with stakeholder goals
/arckit.requirements Create requirements for benefits eligibility chatbot

# Assess compliance
/arckit.ai-playbook Assess AI Playbook compliance for benefits chatbot (HIGH-RISK)
/arckit.tcop Assess TCoP compliance for Beta phase

🔧 What's Changed

Added

OpenAI Codex CLI Support (v0.2.2)

Folder: .codex/ with prompts and setup guide
File: .codex/README.md - Comprehensive Codex CLI setup guide (400+ lines)
File: .codex/prompts/*.md - All 17 ArcKit commands for Codex
File: OPENAI-INTEGRATION-PLAN.md - Integration strategy document
Updated: README.md - Added Codex CLI to supported agents table
Updated: All 7 test repositories with .codex/ support

Stakeholder Analysis (v0.2.1)

Command: /arckit.stakeholders - Comprehensive stakeholder driver analysis
Template: templates/stakeholder-drivers-template.md (400+ lines)
- Power-Interest Grid
- 7 types of drivers (STRATEGIC, OPERATIONAL, FINANCIAL, COMPLIANCE, PERSONAL, RISK, CUSTOMER)
- Driver → Goal → Outcome traceability
- Conflict analysis framework
- Resolution strategies (PRIORITIZE, COMPROMISE, PHASE, INNOVATE)
- RACI matrix
- Engagement plan

Enhanced Requirements Process (v0.2.1)

Updated: /arckit.requirements command
- Now checks for stakeholder analysis first
- Traces requirements back to stakeholder goals
- Identifies and resolves requirement conflicts
- Documents stakeholder management (who won/lost)
Updated: templates/requirements-template.md
- Added "Requirement Conflicts & Resolutions" section
- Added trade-off analysis tables
- Added stakeholder traceability references
- Added 6 common conflict patterns with resolutions

Documentation

Updated: README.md - Corrected workflow order (stakeholders before requirements)
Updated: .claude/COMMANDS.md - Added stakeholder analysis step

Changed

Workflow Order (BREAKING CONCEPT CHANGE)

Old: Principles → Requirements → Design
New: Principles → Stakeholders → Requirements → Design

Impact: Requirements are now informed by stakeholder goals, not invented needs.

Migration:

No code changes required
Recommend running /arckit.stakeholders before /arckit.requirements on new projects
Existing projects can add stakeholder analysis retroactively

🚀 Migration from v0.2.1

For Existing Projects

No breaking changes - all v0.2.1 commands continue to work.

Recommended for new projects:

Run /arckit.stakeholders before /arckit.requirements
When running /arckit.requirements, it will read your stakeholder analysis and:
- Trace requirements back to stakeholder goals
- Identify conflicts from stakeholder conflicts
- Suggest resolutions based on stakeholder power/interest

For OpenAI Codex CLI users:

Pull latest code with .codex/ folder
Set export CODEX_HOME="$(pwd)/.codex"
Start codex --auto
Use /prompts:arckit.* format for commands

For Test Repositories

All 7 test repositories updated with:

.codex/ folder and README
Updated README.md mentioning Codex CLI support
Updated QUICKSTART.md with stakeholder analysis step

📦 Installation

# Install latest version
pip install git+https://github.com/tractorjuice/arc-kit.git

# Or with uv
uv tool install arckit-cli --from git+https://github.com/tractorjuice/arc-kit.git

# Or run without installing
uvx --from git+https://github.com/tractorjuice/arc-kit.git arckit init my-project

# Initialize with Codex CLI support
arckit init my-project --ai codex

🎓 Key Learnings

Why Stakeholders Come Before Requirements

The Problem: Traditional requirements gathering often starts with "what do we need?" This leads to:

❌ Requirements driven by personal opinions, not stakeholder goals
❌ Arbitrary prioritization based on gut feel
❌ Conflicts discovered late when stakeholders review
❌ Poor stakeholder buy-in because their needs weren't understood

The Solution: Start with "who cares and why?"

✅ Requirements trace to real stakeholder goals
✅ Prioritization based on stakeholder power/interest
✅ Conflicts identified early from stakeholder analysis
✅ Strong buy-in because stakeholders see their goals addressed

Example:

Bad (Old Way):

Requirement: "System must have <2 second response time"
Why? "Because fast is good"
Priority? "HIGH because I think it's important"

Good (New Way):

Requirement: "System must have <2 second response time"
Why? "Addresses Product Manager's goal G-3: Improve conversion rate 25%"
Context: "PM's driver D-2: Competitive pressure (STRATEGIC, HIGH intensity)"
Evidence: "PM's research shows 1-second delay = 7% conversion drop"
Priority: MUST (PM is High Power, High Interest stakeholder)
Conflict: None (CISO wants security, no conflict with performance)

🙏 Acknowledgements

This release builds on:

OpenAI Codex CLI - For providing an excellent AI coding agent
Claude Code - For pioneering the AI-assisted architecture workflow
User feedback - For requesting multi-agent support
UK Government frameworks - TCoP, AI Playbook, ATRS (from v0.2.0)

📄 License

MIT License - see LICENSE for details

📈 What's Next (v0.3.0)

Potential future enhancements:

GitHub Actions integration - Automated compliance checks in CI/CD
Dashboard generation - Visual stakeholder maps and traceability charts
Template versioning - Track template evolution over time
Multi-project support - Cross-project principle inheritance
API mode - Programmatic access for custom tooling

Full Changelog: v0.2.1...v0.2.2

v0.2.2 - OpenAI Codex CLI Support & Enhanced Stakeholder Analysis