BUILD DESIGN AGENT — Setup & Upload Checklist

Project Knowledge (uploaded once, lives permanently)

These go into the Claude project's knowledge base during initial setup. You don't re-upload these per client.

1. Project instructions — build-design-agent-project-instructions-v3.md The 10-step design process, sequencing principles, output format, critical rules.

1. System Maturity Model — system-maturity-model-generic.html Defines the four levels (Structure → Streamline → Delegate → Autonomous). Agent needs this to know what a Level 1 build looks like vs. Level 2+.

1. 02-context.md — Advisor defaults (Kathryn Brown, Advisory OS branding). Fallback for any client where a reference data file hasn't been created yet.

Per-Client Uploads (every new conversation)

Upload all of these at the start of each client conversation.

Required (agent cannot run without these)

1. Client Reference Data — client-reference-data-[name].md Correct spellings for company, team, tools. Overrides everything else.

1. Constraint Priority Matrix — constraint-priority-matrix-[name]-[date].md The full matrix output. Agent parses upstream/downstream, synergies, patterns, and tiers from this.

1. Client GPS — Found inside the Client Master Plan or provided as manual input. Position, Direction, Speed. Agent uses this for sequencing and timing decisions.

1. Session transcript — The most recent session where the constraint was discussed. The matrix tells the agent WHAT to solve. The transcript tells the agent HOW the client described the problem — their language, what they've already answered, what their team actually does, what tools they referenced. Without it, the deployment plan will be structurally correct but generic on details.

Recommended (significantly improves output quality)

1. Client Master Plan — client-master-plan-[name].html Full relationship picture. GPS, themes, constraints, engagement history, conversation recaps. If you upload this, you don't need to provide GPS separately.

1. Constraint briefs — If the client has submitted any. Additional evidence the matrix was built from.

Optional (useful but not blocking)

1. Existing Project Plan — If one already exists from a previous run. Agent can refine rather than rebuild from scratch.

1. Existing Client Blueprint — If the client has already been shown a build plan. CRITICAL: agent must align with or flag deviations from what the client has seen.

1. JSON extractions — Relay output from sessions. Structured data for quotes, action items, GPS signals.

Do NOT upload

• Client Roadmap — that's engagement-level, not initiative-level
• AISOP Factory outputs — different agent, different stage
• Other clients' files — one client per conversation

After the agent produces output

Take the deployment plan to the Intelligence Card System project to assemble the Project Plan and Client Blueprint. The Build Design Agent does not produce those documents — it produces the plan they're built from.