the-agency-ai/the-agency/design

Design — A&D Completeness

Drive toward complete Architecture & Design (A&D) using the 1B1 protocol with a completeness checklist. Typically follows /define — the PVR informs architecture.

Usage

• /design — start fresh, auto-detect project
• /design "Project Name" — start fresh for named project
• /design --from <pvr-file> — use PVR as input
• /design --continue <ad-file> — resume work on an existing A&D

Behavior

Step 1: Establish context

If --from specified, read the PVR. If --continue, read the existing A&D. If a PVR exists for this project, load it automatically.

Step 2: Scan for existing work

Glob agency/workstreams/*/ad-*.md for existing A&D files. If found, offer to continue.

Step 3: Build discussion agenda

Map the completeness checklist to discussion items. Skip items already covered.

Step 4: Start discussion

Invoke /discuss via the Skill tool, passing the agenda items.

Step 5: Update A&D progressively

After each item resolves, update the A&D file. Also update the PVR if design decisions reveal requirement changes.

A&D location: agency/workstreams/{workstream}/ad-{workstream}-{slug}-{YYYYMMDD}.md.

Step 6: Present completeness scorecard

A&D Completeness:
 
1. Architecture Overview    ✓ Complete
2. Data Model               ✓ Complete
3. Interfaces               ✓ Complete
4. Dependencies             ✓ Complete
5. Technology Choices       ✓ Complete
6. Trade-offs               ✓ Complete
7. Failure Modes            ~ Partial
8. Security Considerations  ✓ Complete
9. Deployment & Operations  ✓ Complete
10. Open Technical Questions 1 remaining
 
Score: 9/10 complete

Completeness Checklist

1. Architecture Overview — High-level system design, component relationships
2. Data Model — Entities, relationships, storage strategy
3. Interfaces — APIs, protocols, contracts between components
4. Dependencies — External services, libraries, infrastructure
5. Technology Choices — Languages, frameworks, databases, and why
6. Trade-offs — What was considered and rejected, with reasoning
7. Failure Modes — What can go wrong? How does the system degrade?
8. Security Considerations — Auth, data protection, attack surface
9. Deployment & Operations — How it runs, scales, is monitored
10. Open Technical Questions — Unresolved technical items