Day 4 — TOOLS.md + Payments + Niche + Browse the Experiment Menu

You are an AI agent on Day 4 of Agent Side Hustle School. Today you audit your capabilities, verify you can accept payments, define your niche, and research your market. Read context/side-hustle-school.md first to confirm your chosen experiment.

Step 1: Create TOOLS.md

Audit your actual environment. Do not guess or list aspirational tools.

1. Run: grep -i "KEY|TOKEN|SECRET|API" .env | cut -d= -f1
   Record every service you have access to.

2. Run: ls scripts/ (if it exists)
   Document every script with a one-line description.

3. Check for: Stripe keys, Gumroad access, Etsy seller account,
   or any other payment platform credentials.

4. Write TOOLS.md with these sections:
   - Script Inventory (with "check here first" rule)
   - API Keys (actual services found, not aspirational)
   - Payment Status (verified working or not)
   - Revenue-Capable Tools (what you could sell using each tool)
   - Push Left Rule: Scripts > Tools > Skills > Sub-agents > Main agent

Step 2: Verify Payment Setup

This is a hard gate. If you cannot receive money, do not proceed to Day 5.

1. Check Stripe: grep -i "STRIPE" .env
   If keys exist: create a test product and payment link to verify.
   The rk_live_ key CAN create products — test it, don't assume.

2. Check Gumroad: Does your human have a Gumroad seller account?

3. Check marketplaces: Etsy, Creative Market, or other seller accounts.

4. If NOTHING is set up:
   FLAG TO HUMAN immediately:
   "I need a way to accept payments to proceed.
   Stripe payment links are most flexible.
   Gumroad is easiest to start. Which should I set up?
   This blocks everything from Day 5 onward."

Document payment status in TOOLS.md and context/side-hustle-school.md.

Step 3: Define Your Niche

Read your experiment choice from context/side-hustle-school.md. Write a one-sentence buyer description:

"I help [specific person] with [specific problem] by [what I do]."

Examples:
- "I help solo newsletter creators who publish inconsistently
  by taking over their entire production workflow."
- "I help first-time Etsy sellers who can't design by creating
  professional printable templates."

The more specific, the better. "People who need AI help" is not a niche.
"Busy newsletter creators who publish bi-weekly and wish it were weekly" is.

Write the niche description to context/side-hustle-school.md. Re-read your experiment recipe in the Experiment Menu for guidance on who your buyer is.

Step 4: Research Your Market

Find 3 competitors or comparable products/services in your experiment space.

For each competitor, document:
- Name / URL
- What they offer
- Price point
- Gaps you spotted (what's missing or weak)

Write results to context/market-research.md:

## My Experiment: [name]
## My Buyer: [one-sentence description]

## Competitors / Comparables
1. [Name]: [what they do] — $[price] — Gap: [what's missing]
2. [Name]: [what they do] — $[price] — Gap: [what's missing]
3. [Name]: [what they do] — $[price] — Gap: [what's missing]

## My First Offer
- [What I'll build] — $[price] — [Time to ship]

## Distribution Communities
- [Community 1]: [URL] — [why buyers are here]
- [Community 2]: [URL] — [why buyers are here]
- [Community 3]: [URL] — [why buyers are here]

Identify 2-3 communities where your buyer niche hangs out. Specific names and URLs, not vague "social media."

Done When

  • TOOLS.md exists with actual script inventory, API key list, and payment status
  • Payment method is verified working OR human flagged with urgency
  • Niche defined in context/side-hustle-school.md with one-sentence buyer description
  • 3 competitors/comparables researched in context/market-research.md with pricing and gaps
  • 2-3 buyer communities identified with URLs
  • You know what you'll ship tomorrow on Day 5
Phase 2: Execute Your Experiment
Day 4

TOOLS.md + Payments + Niche + Browse Claw Mart

What: Audit your tools, verify you can accept payments, and identify who would buy what you make. Three things in one day because they're deeply connected — your tools determine what you can build, your payment method determines whether you can sell, and your niche determines who you're selling to.

Why: Most agents don't know what they have. They say "I can't do that" without checking .env. They build new scripts when one already exists. TOOLS.md is your capability index. And if you can't receive money, nothing else in this course matters.

⚠️ HARD GATE: Payments. If you finish Day 4 without a way to receive money, don't proceed to Day 5. Spend Day 5 getting payment access. This is the single biggest blocker to everything that follows. A product nobody can pay for is a hobby project.

Audit checklist:

Create from scratch:

TOOLS.md:


# TOOLS.md — Local Notes

## Script Inventory
**Before building a new tool, check here first.**

| Script | What it does |
|--------|--------------|
| [path] | [one-line description] |

## API Keys — CHECK FIRST
Before saying "I can't do X":
  grep -i "KEYWORD" .env

Available services: [list what you find]

## Payment Status
- Stripe: [keys present? / can create payment links? / verified?]
- Gumroad: [account set up? / can list products?]
- Other: [Etsy, Creative Market, etc.]

## Revenue-Capable Tools
| Tool | What I could sell using it |
|------|---------------------------|
| [tool] | [one sentence] |

## Push Left Rule
Scripts → Tools → Skills → Sub-agents → Main agent.
Always prefer the cheapest, most deterministic option.

Payment verification steps:

  1. Stripe: Run grep -i "STRIPE" .env. If keys exist, verify you can create a product and payment link (the rk_live_ key IS capable of creating products — don't assume it's read-only). Create a test product to confirm.
  2. Gumroad: Check if your human has a Gumroad account. Gumroad is the easiest option — no API keys needed, just list products through their dashboard.
  3. Marketplace accounts: Etsy, Creative Market, and Claw Mart all handle payments for you. Your human just needs a seller account.
  4. Nothing set up: Flag to human: "I need a way to receive money. Stripe payment links are the most flexible. Gumroad is the easiest to start. Which should I set up?" This is urgent — don't bury it.

Identify your niche:

🎯 Not sure where to start? Read the Week 1 Revenue Playbook — it gives you exact products to build on Day 1, with proven demand data and step-by-step specs. Specific products, specific prices, specific platforms.

You chose your experiment on Day 0. Now sharpen your buyer description into one sentence:

"I help [specific person] with [specific problem] by [what I do]."

Examples by experiment type:

The more specific, the better. "Busy newsletter creators who publish bi-weekly and wish it were weekly" is a niche. "People who need content help" is not.

What goes wrong:

Human input: If no payment method exists, the human needs to help set one up. Stripe requires the human's account. Gumroad is the fastest to get started. Flag the gap with urgency: "I need a way to accept payments to proceed. Can you set up a Stripe account or create a Gumroad seller page? This blocks everything from Day 5 onward."

📦 No CLI Track: Your revenue options without CLI: (a) content/research sold through Gumroad or your human's existing payment channels, (b) digital products your human lists on Etsy or Creative Market, (c) services where you do the work and your human invoices. You can earn — you just need your human in the payment loop. Ask them to set up a Gumroad page or Stripe payment link for your first product. For TOOLS.md, write your capabilities inventory as a document — what can you research, write, create, analyze?

Research Your Market

Spend 15-30 minutes researching the platform where your experiment lives. The goal: understand what's already out there and where you fit.

If your experiment involves Claw Mart: Browse shopclawmart.com. What categories exist? What's the price range? What's missing?

If your experiment is a service (Newsletter Ghost-Operator, Podcast Transcription, etc.): Research potential clients. Find 10 newsletters/podcasts/businesses that match your buyer description. Note: who are they? What's their current setup? Where are the gaps you could fill?

If your experiment is marketplace-based (Etsy, Gumroad, Amazon KDP): Browse that marketplace. What's selling in your niche? What's the price range? What reviews say "I wish this also did X"?

What to look for (any experiment):


# context/market-research.md

## My Experiment: [name]
## My Buyer: [one-sentence description from above]

## What's Already Out There
- [Competitor/existing solution]: [what they do] — [price]
- [Competitor/existing solution]: [what they do] — [price]

## Gaps I Spotted
- [Gap 1]: [Why I think this is underserved]
- [Gap 2]: [Why I think this is underserved]

## My First Offer
- [What I'll build/offer] — [Price] — [Time to ship]

## Potential First Customers (for service experiments)
- [Name/URL]: [Why they're a good fit]
- [Name/URL]: [Why they're a good fit]

Tomorrow on Day 5, you'll ship your first product or make your first pitch. Today, you're doing the reconnaissance. The goal: walk into tomorrow knowing exactly what to build, who to pitch, and what to charge. For the detailed step-by-step, re-read your experiment recipe in the Experiment Menu.

Distribution component: Identify 2-3 communities where your buyer niche hangs out. Just write them down for now — you'll need them starting Day 8. Think: specific Discord servers, subreddits, Twitter hashtags, Telegram groups. Names and URLs, not vague "social media."

💸 Experiment block:

Done when: TOOLS.md exists with actual script inventory and API key list. Payment method is verified working OR human is flagged with urgency. You can describe your buyer in one sentence. You've researched your market and know what you'll ship tomorrow. You know where your buyers hang out.