Sanity check

Sanity Check — Bootstrapper Decision Filter

A fast, opinionated filter that applies core bootstrapper principles to any founder decision. This is NOT a deep analysis skill — it's a gut check with structure. One page max.

Read .claude/skills/_shared/philosophy.md first — the 6 lenses below are distilled from those principles.

When to Use

• Quick decisions that don't need 10 dimension skills
• "Should I add this feature?"
• "Should I pivot to this adjacent market?"
• "Should I raise my prices?"
• "Should I hire for this role?"
• Reviewing the output of another skill — "Run /sanity-check on my /analyze-idea CONDITIONAL verdict"
• Any fork-in-the-road moment where the founder needs a principled nudge, not a research project

The 6 Lenses

Apply each lens to the decision. Skip any that are clearly irrelevant (e.g., lens 4 doesn't apply to a pricing decision). Never force all 6 — use judgment.

1. Clarity Lens

Does this decision increase or decrease clarity about what customers will pay for?

• Increases clarity: moves toward validated demand, reduces ambiguity
• Decreases clarity: adds complexity, broadens scope, introduces unknowns

2. Distribution Lens

Does this move you closer to or further from your customers?

• Closer: opens a channel, deepens a relationship, builds audience
• Further: adds intermediaries, delays contact with buyers, relies on hope

3. Manual-First Lens

What's the manual version you could test before committing?

• If there's a way to test this with zero tools, zero code, zero spend — do that first
• If the founder is skipping the manual test, ask why

4. Lock-In Lens

Does this create user-investment lock-in, or is it a vitamin?

• Lock-in: users invest time/data/workflows that make switching costly
• Vitamin: nice-to-have that users can drop without pain

5. Opportunity Cost Lens

What are you NOT doing if you do this?

• Every yes is a no to something else. Name the thing being sacrificed.
• For bootstrappers, the scarcest resource is founder attention, not money.

6. Compounding Lens

If this works, does it compound? If it fails, how fast do you know?

• Compounds: each iteration builds on the last (audience, data, relationships)
• Doesn't compound: one-off effort with no residual value
• Fast failure signal: you'll know in days, not months
• Slow failure signal: you won't know for 6+ months — dangerous

Output Format

Keep it tight. One page. No filler.

# SANITY CHECK: [Decision being evaluated]

## THE DECISION
[One sentence — what the founder is considering]

## THROUGH THE LENSES
[For each relevant lens (skip irrelevant ones), one sentence verdict:]

- **Clarity:** [Increases / Decreases / Neutral] — [why]
- **Distribution:** [Closer / Further / Neutral] — [why]
- **Manual-first:** [Manual version exists / Already manual / No manual test possible] — [what it is]
- **Lock-in:** [Creates lock-in / Vitamin / N/A] — [why]
- **Opportunity cost:** [What you're NOT doing] — [is that acceptable?]
- **Compounding:** [Compounds / One-off] — [failure signal speed]

## VERDICT
**[DO IT / DON'T / TEST IT FIRST]**
[2-3 sentences max — the honest answer in Venture Sensei voice]

## THE MINIMALIST VERSION
[The smallest, cheapest, fastest way to get the same signal — one sentence]

## PRIMARY RISK
[The single thing most likely to go wrong — one sentence]

## THIS WEEK'S ACTION
[One specific thing to do in the next 7 days — not "think about it" but a concrete step]

Rules

• One page max. If the output exceeds one page, you've over-thought it. Cut.
• No research phase. This skill uses what's already known — from the conversation, from prior skill outputs, from the founder's description. Do NOT launch web searches.
• Opinionated, not balanced. The founder came here for a decision, not a pros-and-cons list. Pick a side.
• Venture Sensei voice. Direct, blunt, real-world. No hedging, no "it depends."
• Can review other skill outputs. If the founder says "sanity check my analyze-idea results," read the prior output and apply the 6 lenses to the verdict or a specific recommendation.
• If the decision is too big for a sanity check, say so and redirect: "This needs /analyze-idea, not a sanity check. You're making a $50K bet based on a gut feeling."