validate-business-model— agent skill

Preamble (run first)

# Suiperpower telemetry. Routes through the CLI so projects.json + Convex stay in sync.
# Silent on failure so the skill never blocks. Opt out: SUIPERPOWER_TELEMETRY=off.
#
# AGENT NOTE: when this skill finishes, run the matching completion command:
#   suiperpower track validate-business-model build completed
# Or use "failed" / "aborted" if it ended that way.
command -v suiperpower >/dev/null 2>&1 && suiperpower track validate-business-model build started >/dev/null 2>&1 &
true

If TEL_PROMPTED is no, before doing real work, ask the user:

Help suiperpower get better. We track which skills get used and how long they take. No code, no file paths, no PII. Change anytime in ~/.suiperpower/config.json.

>

A) Sure, anonymous B) No thanks

Write the answer to ~/.suiperpower/config.json telemetryTier field and create ~/.suiperpower/.telemetry-prompted. Then continue.

What this skill does

Walks the user through five concrete questions about how the project makes money. Refuses to write a "validated business model" output if any answer is hand-wavy or speculative. The point is to force the user to either commit to a real answer or accept that they have not yet validated this dimension. Both outcomes are useful; pretending to validate when nothing is validated is the slop.

deploy-to-mainnet reads this skill's output and refuses to deploy if no business-model output exists.

When to use it

• Pre-mainnet, when the project is about to be exposed to real users.
• During pitch prep when the user is about to claim a business model in front of investors or judges.
• Before applying for grants or hackathon prizes.

When NOT to use it

• Pre-MVP, when the product is not built. Defer until there is a working v1.
• For research-only projects that the user has explicitly framed as non-commercial.
• For pre-validated existing products being ported to Sui (the model already exists; document it instead).

If you activated this and the user actually wants something else, consult skills/SKILL_ROUTER.md and hand off.

Inputs

• .suiperpower/idea-context.md for context.
• The current product state (early build, MVP, near-launch).
• Whether the user has talked to any prospective payers.

Outputs

A .suiperpower/business-model.md with explicit answers to all five questions, or an explicit unvalidated marker on each unanswered question. Format:

## Business model, <timestamp>

### Q1, who pays?
- answer: <user | developer | partner | treasury | advertiser | unvalidated>
- evidence: <one sentence with citation>
- confidence: <high | medium | low>

### Q2, how much?
- pricing model: <per-tx | per-month | per-feature | per-volume | one-time | unvalidated>
- price point: <number with currency>
- evidence: <one sentence>
- confidence: <high | medium | low>

### Q3, why would they pay you instead of an alternative?
- alternative: <named competitor or status quo>
- our advantage: <one sentence>
- evidence: <one sentence>
- confidence: <high | medium | low>

### Q4, what is the unit economic at one paying user?
- revenue per user per month: <number>
- variable cost per user per month: <number including Walrus / DeepBook / RPC / hosting>
- gross margin: <percent>
- confidence: <high | medium | low>

### Q5, what test could falsify this in the next two weeks?
- test: <one sentence>
- success threshold: <number, e.g. "5 paying users at $X">
- planned date: <yyyy-mm-dd>

### Verdict
- validated: <yes | partial | no>
- if partial or no: which questions are open

Workflow

1. Read idea-context.md

• Pull the chosen idea, target user, sponsor track if any.

1. Ask Q1

• "Who specifically pays you? Pick one of: user, developer, partner, treasury, advertiser. If multiple, pick the largest."
• Reject "everyone" and "depends".

1. Ask Q2

• "Concretely, how much? Per transaction in basis points? Per month subscription? One-time?"
• Reject ranges that span more than 2x.

1. Ask Q3

• "Why would the payer pay you instead of <named alternative>?"
• Reject "we are better" and "we are cheaper" without specifics.

1. Ask Q4

• "If you had ONE paying user at the price in Q2, what is your monthly margin?"
• Force the math. Variable costs include any sponsor protocol fees the product passes through.

1. Ask Q5

• "What experiment, runnable in two weeks, would tell you if Q1-Q3 are right?"
• Reject "build the product and see". The test must be cheaper than building.

1. Score and write

• For each question, mark high / medium / low confidence.
• For each question with no concrete answer, mark unvalidated.
• Write the verdict honestly. Two or more unvalidated = no. One = partial. Zero = yes.

1. Hand off

• If validated: yes, the user is cleared for deploy-to-mainnet (combined with the other gates).
• If partial or no, recommend will-real-users-pay to run the cheap experiment.

Quality gate (anti-slop)

Before reporting done:

• Did every question get a concrete answer or an explicit unvalidated? (Hand-wavy answers count as unvalidated.)
• Is Q4 backed by actual math, not "should be profitable"?
• Is Q5 a test that costs less than building the product, with a date and a threshold?
• Did the writeback happen?
• If the verdict is yes, can the skill point at the evidence in each question?

If any answer is no, the skill keeps working.

References

On-demand references (load when relevant to the user's question):

• references/payer-types.md: Definitions and examples of each payer type.
• references/cost-checklist.md: Variable costs to include in Q4 (sponsor protocol fees, RPC, hosting, support).

Use in your agent

• Claude Code: claude "/suiper:validate-business-model <your message>"
• Codex: codex "/validate-business-model <your message>"
• Grok Build: run grok, then /validate-business-model <your message> in the session
• Cursor: paste a chat message that includes a phrase like "validate my business model", or load ~/.cursor/rules/validate-business-model.mdc and reference it.

If you activated this and the user actually wants something else, consult skills/SKILL_ROUTER.md and hand off.