# Mvp Definition

> >-

- **Type:** Skill
- **Install:** `agentstack add skill-stanislavnianko-product-discovery-claude-skills-mvp-definition`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [stanislavnianko](https://agentstack.voostack.com/s/stanislavnianko)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [stanislavnianko](https://github.com/stanislavnianko)
- **Source:** https://github.com/stanislavnianko/product-discovery-claude-skills/tree/main/plugins/discovery-phase/skills/mvp-definition

## Install

```sh
agentstack add skill-stanislavnianko-product-discovery-claude-skills-mvp-definition
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# MVP Definition

> Part of the **discovery-phase** skill pack · `scoping` group · reads `discovery-context.md` (run `profile-builder` first if missing).

A clarifier skill — runs when "what is the MVP, exactly?" comes up. Distinguishes three levels: PoC (proves a tech assumption), MVP (proves a market assumption), v1 (delivers value at scale). Outputs an explicit comparison so the client and agency align before scope hardens.

## When to run this skill

- Client and agency disagree on what "MVP" means
- The team is conflating "stripped-down product" with "learning experiment"
- After `feature-scoping` to validate the scope is actually MVP-shaped
- During pre-sale conversations to align expectations

## Step 1 — Read context

Read `discovery-context.md` and `risk-assumption-map.md` (top-3 assumptions are what the MVP must validate). Optionally read `scope-doc.md` (reinforces MVP framing).

If `discovery-context.md` is missing, ask the BA inline: "client stage — pre-launch / early traction / scaling / mature?" — tag the output `[ASSUMED]`. If `risk-assumption-map.md` is missing, ask: "name 1–3 assumptions to retire in one sentence each" — or proceed tagged `[NO-RISK-FRAME]` (MVP becomes feature-driven, not learning-driven). Never block; recommend `profile-builder` / `risk-assumption-mapping` for high-stakes work.

## Step 2 — Three-level table

Build the comparison.

| Level | Goal | Audience | Success criterion | Tear-down? | Lifespan |
|---|---|---|---|---|---|
| **PoC** | Retire 1-3 specific assumptions | Internal + maybe 5 friendlies | Assumption resolved (yes/no) | Yes, throwaway | Days-weeks |
| **MVP** | Validate market hypothesis | Real users (limited segment) | Leading metric crosses pre-committed threshold | Maybe — could harden if MVP wins | Weeks-months |
| **v1** | Deliver value at scale | Full target market | Lagging business metric (revenue, retention, NPS) | No | Open-ended |

Fill the rows specific to this engagement.

## Step 3 — State which level applies

Most discovery cycles end with one of:

- "We need a PoC first" → recommends `feasibility-spike` + `prototype-plan` from validation group
- "We're ready for MVP" → use `feature-scoping` + `proposal`/`sow-draft` 
- "v1 is appropriate" → unusual at end of discovery; re-check; usually means discovery validated existing solution

## Step 4 — Pre-committed MVP success criteria (if MVP level applies)

Two numbers. Without these, the MVP isn't an MVP — it's just a build.

- **Leading (during MVP window):** behavior that says "users want this"
- **Lagging (4-8 weeks post-launch):** outcome that says "this changed something"

## Step 5 — MVP exclusion list

What the MVP does NOT include — even though it'd be tempting to add. This is different from `scope-doc.md` "explicitly deferred":

- scope-doc deferred = "later release"
- MVP exclusion = "we'd be wasting money to include this in the *learning experiment*"

Examples: full UI polish, edge-case error handling, scale-grade infra, every integration, multi-language support, full a11y compliance.

## Step 6 — Tear-down decision

State up front: if the MVP **fails** its success criterion, what happens?

- Throwaway, return to pre-MVP state
- Keep as evidence artifact, don't operate
- Hand off to client for them to maintain (rare)

If the MVP **succeeds**, what's the path to v1?

- Harden in place
- Rebuild clean from scratch
- Hand off architecture decisions to delivery team

## Output

`./discovery/mvp-definition.md` per `./template.md`.

Append to `_log.md`: `[mvp-definition | YYYY-MM-DD] level: ; threshold: ; teardown: `.

## Anti-patterns

- **MVP as "v1 minus features".** That's not an MVP — that's a delayed v1. MVP is a *learning vehicle*.
- **No success threshold pre-committed.** Without it, every MVP "succeeds" on whichever metric looks best post-hoc.
- **Conflating PoC and MVP.** PoC retires tech risk for the team; MVP retires market risk against real users. Different audiences, different bars.
- **Open-ended tear-down.** "We'll see how it goes" — that's how MVPs become unmaintained zombies. Decide upfront.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [stanislavnianko](https://github.com/stanislavnianko)
- **Source:** [stanislavnianko/product-discovery-claude-skills](https://github.com/stanislavnianko/product-discovery-claude-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-stanislavnianko-product-discovery-claude-skills-mvp-definition
- Seller: https://agentstack.voostack.com/s/stanislavnianko
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.