Discovery ProcessSAFE

Product Management skills framework built on battle-tested methods for Claude Code, Cowork, Codex, and AI agents.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: discovery-process
argument-hint: "[problem hypothesis]"
description: Run a full discovery cycle from problem hypothesis to validated solution. Use when a team needs a structured path through framing, interviews, synthesis, and experiments.
intent: >-
  Guide product managers through a complete discovery cycle—from initial problem hypothesis to validated solution—by orchestrating problem framing, customer interviews, synthesis, and experimentation skills into a structured process. Use this to systematically explore problem spaces, validate assumptions, and build confidence before committing to full development—avoiding "build it and they will come" syndrome and ensuring you're solving real customer problems.
type: workflow
theme: discovery-research
best_for:
  - "Running a full discovery cycle from hypothesis to validated solution"
  - "Investigating a retention or churn problem systematically"
  - "Setting up continuous discovery as an ongoing practice"
scenarios:
  - "I have a hypothesis that B2B customers struggle with onboarding and want to validate it before building anything"
  - "Our activation rate dropped 15% this quarter and I need to run discovery to find out why"
estimated_time: "30-60 min"
---


## Purpose
Guide product managers through a complete discovery cycle—from initial problem hypothesis to validated solution—by orchestrating problem framing, customer interviews, synthesis, and experimentation skills into a structured process. Use this to systematically explore problem spaces, validate assumptions, and build confidence before committing to full development—avoiding "build it and they will come" syndrome and ensuring you're solving real customer problems.

This is not a one-time research project—it's a continuous discovery practice that runs in parallel with delivery, typically 1-2 discovery cycles per quarter.

## Input

**Works best with:** Your starting problem hypothesis — even a rough one.
**Also useful:** Prior research, customer access, timeline, and what decision the discovery must inform.

Anything supplied with the invocation itself — text after the skill name, a pasted context dump, or an appended `ARGUMENTS:` line — counts as answers already given. Use it and skip whatever it covers; don't re-ask.

**Arriving empty-handed? That works too.** The workflow starts at problem framing and helps you construct the hypothesis first.

**Example invocation:** `Run discovery on this hypothesis: SMB admins abandon onboarding because the data-import step requires IT help they don't have.`

## Key Concepts

### What is the Discovery Process?

The discovery process (Teresa Torres, Marty Cagan) is a structured approach to exploring problem spaces and validating solutions before building. It consists of:

1. **Frame the Problem** — Define what you're investigating and why
2. **Conduct Research** — Gather qualitative and quantitative evidence
3. **Synthesize Insights** — Identify patterns, pain points, and opportunities
4. **Generate Solutions** — Explore multiple solution options
5. **Validate Solutions** — Test assumptions through experiments
6. **Decide & Document** — Commit to build, pivot, or kill

### Why This Works
- **De-risks product decisions:** Tests assumptions before expensive builds
- **Customer-centric:** Grounds decisions in real customer problems, not internal opinions
- **Iterative:** Builds confidence progressively through small experiments
- **Fast learning:** Discovers "no-go" signals early, saves wasted effort

### Anti-Patterns (What This Is NOT)
- **Not waterfall research:** Discovery runs continuously, not once before dev
- **Not user testing:** Discovery validates problems; testing validates solutions
- **Not a substitute for shipping:** Discovery informs delivery, doesn't replace it

### When to Use This
- Exploring new product/feature areas
- Investigating retention or churn problems
- Validating strategic initiatives before roadmap commitment
- Continuous discovery (weekly customer touchpoints)

### When NOT to Use This
- For well-understood problems (move to execution)
- When stakeholders have already committed to a solution (address alignment first)
- For tactical bug fixes or technical debt (no discovery needed)

---

### Facilitation Source of Truth

When running this workflow as a guided conversation, use [`workshop-facilitation`](../workshop-facilitation/SKILL.md) as the interaction protocol.

It defines:
- session heads-up + entry mode (Guided, Context dump, Best guess)
- one-question turns with plain-language prompts
- progress labels (for example, Context Qx/8 and Scoring Qx/5)
- interruption handling and pause/resume behavior
- numbered recommendations at decision points
- quick-select numbered response options for regular questions (include `Other (specify)` when useful)

This file defines the workflow sequence and domain-specific outputs. If there is a conflict, follow this file's workflow logic.

## Application

Use `template.md` for the full fill-in structure.

This workflow orchestrates **6 phases** over **2-4 weeks**, using multiple component and interactive skills.

---

## Phase 1: Frame the Problem (Day 1-2)

**Goal:** Define what you're investigating, who's affected, and success criteria.

### Activities

**1. Run Problem Framing Canvas**
- **Use:** `skills/problem-framing-canvas/SKILL.md` (interactive - MITRE)
- **Participants:** PM, design, engineering lead
- **Duration:** 120 minutes
- **Output:** Problem statement + "How Might We" question

**2. Create Formal Problem Statement**
- **Use:** `skills/problem-statement/SKILL.md` (component)
- **Participants:** PM
- **Duration:** 30 minutes
- **Output:** Structured problem statement with hypothesis

**3. Define Proto-Personas (If Needed)**
- **Use:** `skills/proto-persona/SKILL.md` (component)
- **When:** If target customer segment is unclear
- **Duration:** 60 minutes
- **Output:** Hypothesis-driven personas

**4. Map Jobs-to-be-Done (If Needed)**
- **Use:** `skills/jo

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.