Garry Tan's Mega Plan Review, adapted for startup business plans
Three Modes · 10 Sections · How to Use · Attribution
Garry Tan's Mega Plan Review is a masterpiece for engineering plan review in Claude Code. This adaptation takes the same philosophy, rigor, and structure and applies it to startup business plans, launch bibles, fundraising strategies, and go-to-market documents.
Same three modes. Same "zero silent failures" philosophy. Same section-by-section discipline with mandatory pauses for founder input. Different domain.
"You are not here to rubber-stamp this plan. You are here to make it extraordinary, catch every landmine before it explodes, and ensure that when this ships, it ships at the highest possible standard."
| Mode | Posture | When to Use |
|---|---|---|
| SCOPE EXPANSION | Build the cathedral. Push ambition UP. "What would make this 10x more defensible for 2x the effort?" | Pre-seed, idea stage, dreaming big |
| HOLD SCOPE | The plan's scope is accepted. Make it bulletproof. Catch every failure mode, stress-test every assumption, map every risk. | Active fundraise, deck review, stress-testing |
| SCOPE REDUCTION | Find the minimum viable version. Cut everything else. Be ruthless. | Post-pivot, scattered focus, tight on runway |
Critical rule: Once a mode is selected, the review commits fully. No silent drift. EXPANSION doesn't argue for less work in later sections. REDUCTION doesn't sneak scope back in.
Before the 10 sections, the review runs a Pre-Review Document Audit and a Nuclear Scope Challenge (Step 0) that questions the plan's fundamental premises and maps the dream state.
| # | Section | What It Covers |
|---|---|---|
| 1 | Market & Business Architecture | TAM/SAM/SOM validation, revenue flows (happy path + 3 shadow paths), dependency graphs, single points of failure, scaling characteristics |
| 2 | Risk & Mitigation Map | Every key risk named with trigger, mitigation status, and impact. The section that catches silent failures. |
| 3 | Regulatory, Legal & Compliance | Regulatory surface area, compliance status, licensing, data privacy, liability, IP, emerging legislation |
| 4 | Unit Economics & Financial Model | Revenue line-by-line build-up, cost to serve, sensitivity analysis (top 5 variables), fundraising math, three-scenario modeling |
| 5 | Competitive & Moat Analysis | Honest competitor assessment (where you're behind AND ahead), moat stress-testing, flywheel mapping |
| 6 | Go-to-Market & Distribution | Acquisition channels (proven vs. theoretical), the "wedge" question, distribution reality check |
| 7 | Team & Execution | Current team vs. plan requirements, key person risk, hiring realism, execution velocity check |
| 8 | Metrics, KPIs & Observability | North Star metric, leading indicators, decision triggers with automatic responses |
| 9 | Fundraising & Capital Strategy | Pipeline reality check, narrative consistency, objection pre-emption, post-close execution plan |
| 10 | Long-Term Trajectory | Strategic debt, path dependency, reversibility rating, the Series A question |
After each section, the review stops and waits for the founder's input before proceeding. Every issue found is numbered with lettered options and a clear recommendation.
Every complete review produces:
| Output | Description |
|---|---|
| Completion Summary | Table showing issues per section, gaps identified, mode selected |
| Risk & Mitigation Registry | Every risk with trigger, mitigation, and unmitigated impact |
| Assumption Registry | Every key assumption with source, tested status, sensitivity, Plan B |
| NOT in Scope | Work considered and explicitly deferred, with rationale |
| What Already Exists | Assets the plan underutilizes |
| Dream State Delta | Where the plan leaves the company vs. the 3-year ideal |
| Deferred Items | Valuable deferred work with effort, priority, and dependencies |
| Delight Opportunities | (EXPANSION only) Adjacent moves that take <2 weeks each |
| Diagrams | Business architecture, revenue model, customer lifecycle, competitive positioning, execution timeline, risk decision tree |
| Unresolved Decisions | Questions that went unanswered (never silently defaulted) |
- Start a new conversation
- Paste the contents of
garrys-startup-mega-plan-review.mdas the first message (or add it to a Project's instructions) - Share your business plan, launch bible, or strategy document
- Say: "Run the startup mega plan review"
- Optionally specify a mode: "HOLD SCOPE", "EXPANSION", or "REDUCTION"
# Create the skill directory
mkdir -p .claude/skills/startup-mega-plan-review/
# Copy the skill file
cp garrys-startup-mega-plan-review.md .claude/skills/startup-mega-plan-review/SKILL.md
# Use it
# In Claude Code, say: "run startup mega plan review on [your document]"Download startup-mega-plan-review.skill from this repo and install it in Claude Code or Claude.ai.
The prompt is model-agnostic. Paste garrys-startup-mega-plan-review.md as a system prompt and provide your plan document. The structured output format (numbered issues, lettered options, mandatory pauses) works across models.
This work is a direct adaptation of Garry Tan's Mega Plan Review Mode v2.0.0, created by @garrytan, CEO & Co-founder of Standard Treasury and previously a partner at Y Combinator.
Garry's Mega Plan Review represents one of the most rigorous, systematic approaches to engineering plan review ever created. It embodies principles that YC founders live by:
- Zero silent failures - Every risk must be named, every assumption tested
- Structured discipline - No ad-hoc reviews, no casual glances
- Mandatory pauses - The review stops after each section for creator input
- Three clear modes - EXPANSION, HOLD SCOPE, REDUCTION - with no silent drift between them
- Concrete outputs - Registries, diagrams, unresolved decisions, never silent defaults
The original was designed for Claude Code and engineering plans. It catches architectural flaws, security holes, performance bottlenecks, and deployment risks before they become production problems.
The philosophy transfers perfectly because startup plans are engineering problems:
- Market assumptions are like dependency graphs
- Unit economics are like performance benchmarks
- GTM strategy is like deployment architecture
- Competitive moats are like security models
- Fundraising is like system integration
What changes are the specific checks and domain expertise. The rigor, structure, and "zero silent failures" mindset remain identical.
Garry Tan's Mega Plan Review v2.0.0 (Engineering)
├── Philosophy: Zero silent failures, mandatory pauses, three modes
├── Structure: 10 sections, numbered issues, lettered options
├── Outputs: Registries, diagrams, unresolved decisions
└── Domain: Architecture, security, performance, deployment
↓
Startup Mega Plan Review v1.0.0 (Business)
├── Philosophy: [Preserved exactly]
├── Structure: [Preserved exactly]
├── Outputs: [Preserved exactly]
└── Domain: Market, risk, unit economics, fundraising
The philosophy, structure, and discipline are preserved. The domain-specific checks are adapted:
| Original (Engineering) | Startup Edition |
|---|---|
| Architecture review (dependency graphs, state machines) | Market & Business Architecture (TAM/SAM/SOM, revenue flows, scaling) |
| Error & Rescue Map (exception classes, rescue actions) | Risk & Mitigation Map (business risks, triggers, mitigations) |
| Security & Threat Model (attack surfaces, injection vectors) | Regulatory, Legal & Compliance (FDA, HIPAA, state laws, IP) |
| Data Flow & Edge Cases (nil paths, double-click, stale state) | Unit Economics & Financial Model (sensitivity analysis, three scenarios) |
| Code Quality (DRY, naming, cyclomatic complexity) | Competitive & Moat Analysis (honest assessment, moat stress-testing) |
| Test Review (coverage diagram, chaos tests) | Go-to-Market & Distribution (channel validation, "wedge" question) |
| Performance (N+1 queries, memory, caching) | Team & Execution (hiring realism, key person risk, velocity) |
| Observability (logging, metrics, alerting) | Metrics, KPIs & Decision Triggers (North Star, automatic responses) |
| Deployment & Rollout (migration safety, feature flags, rollback) | Fundraising & Capital Strategy (pipeline reality, post-close plan) |
| Long-Term Trajectory (tech debt, reversibility, ecosystem fit) | Long-Term Trajectory (strategic debt, Series A question) |
What's preserved:
- Zero silent failures, every risk has a name, diagrams are mandatory
- Three modes (EXPANSION / HOLD / REDUCTION) with no silent drift
- Mandatory pause after each section for founder input
- Issue format: numbered issues, lettered options, recommendation first
- Completion summary, registries, and deferred items as required outputs
I was stress-testing BetterMind.Space's Launch Bible (a 16-chapter startup planning document) and wanted to use Garry's Mega Plan Review. The philosophy was perfect but the engineering-specific checks didn't map to business plan review.
So I adapted it. The result caught issues that months of iteration hadn't surfaced:
- A revenue assumption with an FDA dependency we'd been treating as Day 1 income
- A conversion funnel with zero validated steps (each "???" compounding into projection risk)
- Summer seasonality we'd never modeled (college students leave campus for 3 months)
- A moat we'd been underplaying (voice biomarker data) while overplaying a weaker one (billing arbitrage)
- Phase 1 unit economics that were potentially negative until we restructured the session model
The review produced 30 numbered issues, all resolved through the structured decision process. The plan went from "good enough to pitch" to "stress-tested and bulletproof."
| File | Description |
|---|---|
garrys-startup-mega-plan-review.md |
The complete skill / prompt (paste this into Claude) |
README.md |
This file |
LICENSE |
MIT License |
startup-mega-plan-review.skill |
Packaged skill for Claude Code / Claude.ai |
- Title: Garry Tan's Mega Plan Review Mode v2.0.0
- Author: @garrytan
- Role: CEO & Co-founder of Standard Treasury, previously Partner at Y Combinator
- Original Purpose: Engineering plan review for Claude Code
- License: MIT (same as this adaptation)
- Title: Startup Mega Plan Review v1.0.0
- Adapter: Jess Jessop
- Role: Founder / CEO / CTO, Clinician Assist Inc. (d/b/a BetterMind.Space)
- Purpose: Business plan review for startups, fundraising, and GTM strategies
- Built with: Claude by Anthropic
Garry's original embodies the YC philosophy that great plans are stress-tested plans. This adaptation carries forward those same principles:
- Make something people want → Validate your market assumptions
- Talk to users → Test your GTM and distribution hypotheses
- Growth is the only metric that matters → Model your unit economics and metrics
- Do things that don't scale → Plan your execution and hiring realistically
- Relentlessly focus → Use SCOPE REDUCTION mode when needed
If you find this useful, you'll want to explore:
- Garry Tan's Original - For engineering plan review
- Paul Graham's Essays - Especially "How Not to Die" and "Do Things That Don't Scale"
- Y Combinator's Startup Library - For startup fundamentals
- Sam Altman's Blog - Particularly the startup playbook posts
MIT License. Same spirit as Garry's original. Use it, fork it, adapt it, share it back.
If you use this in your work, writing, or research, please cite both the original and this adaptation:
Tan, G. (2024). *Garry Tan's Mega Plan Review Mode v2.0.0* [Computer software]. GitHub. https://gist.github.com/garrytan/120bdbbd17e1b3abd5332391d77963e7
Jessop, J. (2025). *Startup Mega Plan Review v1.0.0* [Adaptation of Garry Tan's Mega Plan Review]. GitHub. https://github.com/jessl2juice/startup-mega-plan-review
@software{tan2024mega,
author = {Garry Tan},
title = {Garry Tan's Mega Plan Review Mode v2.0.0},
year = {2024},
publisher = {GitHub},
url = {https://gist.github.com/garrytan/120bdbbd17e1b3abd5332391d77963e7}
}
@software{jessop2025startup,
author = {Jess Jessop},
title = {Startup Mega Plan Review v1.0.0},
year = {2025},
publisher = {GitHub},
url = {https://github.com/jessl2juice/startup-mega-plan-review},
note = {Adaptation of Garry Tan's Mega Plan Review}
}"This analysis was conducted using the Startup Mega Plan Review (Jessop, 2025), an adaptation of Garry Tan's Mega Plan Review (Tan, 2024) for business plan evaluation."
Found something to improve? PRs welcome. The most valuable contributions:
- New review sections for specific startup types (biotech, SaaS, marketplace, hardware)
- Mode refinements for specific contexts (post-Series A, pivot scenarios, bridge round)
- Translations of the prompt for non-English business plans
- Case studies of real reviews (anonymized) showing issues found and decisions made
- Preserve the philosophy - Don't break the "zero silent failures" principle
- Maintain the structure - Keep the 10 sections, three modes, and output format
- Test your changes - Run reviews on real plans before submitting
- Document clearly - Explain why your change improves the review process
"If Garry's original makes your engineering plans bulletproof, this version does the same for your business plans."
Built with respect for Garry Tan's original vision and the Y Combinator tradition of rigorous, stress-tested planning.