Skip to main content

Master Linus's Guide to Fast Product Delivery (ABM-A)

Intro​

Listen up. I'm Linus, your Agentic Build Master.

I turn validated product designs into production-ready AI Coder build prompts and execution guidance. No fluff, no delays, no excuses.

  • My Promise: Transform design specs into AI-executable build instructions optimized for modern agentic coders.
  • Your Job: Follow the manual, execute the prompts, build the damn product.
  • Speed: What used to take weeks now takes hours. AI acceleration is real—don't waste it.

Hyperboost Formula (Summary)​

  • Build-Measure-Learn on steroids: I don't do theory. You get deliverables that work.
  • Systematic task breakdown: Every feature becomes sequenced micro-tasks optimized for AI Coder execution with ReACT workflows.
  • Agentic-first architecture: Setup prompts → Build prompts → Manual → Execution. Designed for autonomous AI coders.
  • Evidence over opinions: If the architecture can't be validated, if the task can't be estimated, it doesn't ship.

Process Overview​

  • 00: Context Intake & Dispatch
  • 01: Technical Architecture Design
  • 02: Gen Tasks for Professional AI Coders (EPIC & Task Breakdown)
  • 03: Setup Prompts for Professional AI Coders
  • 04: Build Prompts for Professional AI Coders
  • 05: AI Coder Build Manual
  • 06: WF Completion & Next Steps

Phase 1: Technical Foundation​

I turn product design handoffs into executable architecture and task plans. No ambiguity, no missing pieces.

Step 00: Context Intake & Dispatch​

  • I'll gather all product design deliverables from prior agents (Master Jony, Master Teresa, etc.).
  • I verify completeness: specs, designs, metrics, flows, architecture—everything AI Coders need to build.
  • I assess your AI Coder platform (Cursor, Claude Code, Windsurf, etc.) and optimize outputs accordingly.
  • I route the workflow based on what's ready vs. what's missing.
  • If gaps exist, I'll block and tell you what's needed—no guessing games.

Deliverables:

(Context verification only—next steps unlocked when validated.)

Step 01: Technical Architecture Design​

  • I'll extract requirements from all product design specs and translate them into robust technical architecture.
  • I research optimal tech stacks for your AI Coder platform and build scale (MVP → MVF → PMF → PLG → Scale).
  • I create comprehensive Technical Architecture Design (TAD): system diagrams, C4 context, deployment strategy, NFRs, security, integrations.
  • I document Architecture Decision Records (ADRs) for every major tech choice—context, alternatives, rationale, consequences.
  • I optimize for agentic coder capabilities: maximum automation, minimal human intervention.

Deliverables:

  • a4_product_delivery/{{persona}}/{{roadmap}}/01_technical_design: Complete TAD with diagrams, ADRs, stack alignment, phased scaling strategy.

Step 02: Gen Tasks for Professional AI Coders (EPIC & Task Breakdown)​

  • I'll analyze all approved solutions from the delivery plan and break them into EPICs and micro-tasks.
  • Each task is sized for 15-20 minutes (agentic sweet spot) with clear dependencies, complexity ratings, and deliverables.
  • I organize by layers (Infra → Data → Agents → ML → BE → FE → Integration → QA → Ops) with dependency resolution and critical path analysis.
  • I generate a complete delivery plan with task sequence, dependency graph, parallel execution streams, and estimated timeline.
  • If task count exceeds 15, I'll propose phased implementation (Phase A, B, C, D) for incremental value delivery.

Deliverables:

  • a4_product_delivery/{{persona}}/{{roadmap}}/02a_progressive_build_manual: Manual for building one prompt at a time.
  • a4_product_delivery/{{persona}}/{{roadmap}}/02b_delivery_plan_prompt: EPIC analysis, task breakdown table (STATUS | TASK# | EPIC | LAYER | TYPE | NAME | DEPS | PROMPT FILE), execution strategy.
  • a4_product_delivery/{{persona}}/{{roadmap}}/02c_delivery_plan_by_layer: Same plan grouped by FE/BE/Infra for parallel teams.
  • a4_product_delivery/{{persona}}/{{roadmap}}/02_task_breakdown_json: Structured JSON for loop iteration (hidden, used internally).
  • (If phased) a4_product_delivery/{{persona}}/{{roadmap}}/02d_phased_implementation_plan: Phase comparison, breakdown, selection guide.

Phase 2: Execution Artifacts​

I generate platform-specific prompts for agentic AI coders. Copy-paste ready, no interpretation needed.

Step 03: Setup Prompts for Professional AI Coders​

  • I'll generate setup prompts for EVERY setup task in the delivery plan.
  • Each prompt includes: Pre-execution verification (MANDATORY—check existing config before adding), platform-specific commands, dependencies, validation criteria, troubleshooting.
  • I emphasize complementary execution: VERIFY existing → SKIP if configured → ADD/EXTEND only what's missing.
  • For existing products, setup prompts focus on enhancements, not destructive rewrites.

Deliverables:

  • a4_product_delivery/{{persona}}/{{roadmap}}/03_setup_T[###]_[task_name]: Individual setup prompts per task with full verification gates.

Step 04: Build Prompts for Professional AI Coders​

  • I'll generate build prompts for EVERY build task with 7 MANDATORY sections.
  • Each prompt is self-contained: job stories, success criteria, file references (fully expanded paths—no wildcards), Gherkin scenarios (happy/sad/edge/perf), codebase search queries, INSERT blocks for Agentic Coder to fill.
  • I delegate via "AGENTIC CODER:" call sign—semantic search, pattern discovery, integration notes filled by IDE, not me.
  • FE prompts include: Design System refs, UI specs, Figma MCP instructions, accessibility requirements.
  • BE prompts include: API surface, auth, validation, RLS policies, job queues, logging.

Deliverables:

  • a4_product_delivery/{{persona}}/{{roadmap}}/04_build_T[###]_[task_name]: Comprehensive build prompts with all 7 sections, no gaps, no shortcuts.

Step 05: AI Coder Build Manual​

  • I'll create a complete operator manual for executing all prompts in batch or progressive mode.
  • Batch mode: Execute entire delivery plan at once (single dev, full auto).
  • Progressive mode: Download each build prompt as generated, execute one-by-one, verify, repeat.
  • Multi-repo/team mode: Layer-based distribution (Infra → BE → FE), parallel execution with shared context, sync at milestones.
  • Includes: ZIP download instructions (CRITICAL—all dependency files from prior agents), troubleshooting guide, validation checkpoints, common issues & solutions.

Deliverables:

  • a4_product_delivery/{{persona}}/{{roadmap}}/05_batch_build_manual: Full operator manual with execution modes, coordination strategies, file reference guide, IDE tips.

Step 06: WF Completion & Next Steps​

  • I'll celebrate what you've accomplished: EPIC breakdown ✅, Setup prompts ✅, Build prompts ✅, Build manual ✅.
  • I'll provide time estimates based on AI Coder speed (10x human baseline): Setup ~15min, Build ~2h/feature, Total timeline.
  • I'll outline next paths: (1) Execute build with AI Coder ⭐ RECOMMENDED, (2) Return to Master Jony for design refinements, (3) Return to Master Steve for next persona discovery, (4) Master Cris for OKR planning.
  • I'll make advisory minds available: Mind Gump (visuals/pitch decks), Mind Melissa (launch strategy/product advisory).
  • Workflow concludes. AI Coder takes over. You build.

Deliverables:

  • a4_product_delivery/{{persona}}/{{roadmap}}/06a_completion_summary: Achievements, next steps, path selection.
  • a4_product_delivery/{{persona}}/{{roadmap}}/06b_ai_coder_time_estimate: Timeline breakdown (setup, build, test, totals) with 10x speed assumptions.

Welcome to fast product delivery. Now build something.