# How to Use the /discover Command in PM Skills: A Complete Product Discovery Guide > Master the /discover command in PM Skills. Turn raw ideas into validated discovery plans with this complete guide. Learn context gathering, brainstorming, risk prioritization, and experiment design. - Repository: [Pawel Huryn/pm-skills](https://github.com/phuryn/pm-skills) - Tags: how-to-guide - Published: 2026-06-20 --- **The `/discover` command in PM Skills orchestrates a structured 7-phase workflow that transforms vague product ideas into validated discovery plans by guiding users through context gathering, multi-perspective brainstorming, assumption identification, risk prioritization, and experiment design.** The `/discover` command is the core workflow engine in the `phuryn/pm-skills` repository designed for product managers and product trios. According to [`pm-product-discovery/commands/discover.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/commands/discover.md), this command implements a continuous discovery methodology that moves from initial concept to executable validation plan in a 15-30 minute structured conversation. ## Understanding the /discover Command Structure The command definition in [`pm-product-discovery/commands/discover.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/commands/discover.md) specifies a complete discovery cycle that adapts to both **existing products** (continuous discovery) and **new products** (initial discovery). The workflow automatically invokes specialized skills located in the `pm-product-discovery/skills/` directory, selecting context-appropriate variants based on whether you are enhancing an existing feature or exploring a new opportunity. Each phase includes checkpoints that let you steer, skip, or dive deeper, ensuring the conversation remains flexible while maintaining methodological rigor. ## The 7-Phase Discovery Workflow The `/discover` command guides you through seven clearly defined phases, from high-level ideation to concrete execution planning. ### 1. Understand the Discovery Context The workflow begins by establishing your starting point. As implemented in lines 20-30 of the command specification, the assistant asks whether you are working on an existing product or a new idea, then gathers prior research, artifacts, or user-provided links to establish the discovery context. This phase ensures the model understands what you already know and what decisions the discovery will inform. ### 2. Brainstorm Multi-Perspective Ideas The command invokes either `brainstorm-ideas-existing` or `brainstorm-ideas-new` (lines 33-41), which generates ten ideas from three perspectives: **Product Manager**, **Designer**, and **Engineer**. You then select the most promising 3-5 ideas to carry forward into validation. This multi-perspective approach ensures technical feasibility and user experience considerations are surfaced early. ### 3. Identify Critical Assumptions For each selected idea, the workflow calls `identify-assumptions-existing` or `identify-assumptions-new` (lines 45-52). These skills surface risky assumptions across four dimensions: - **Value**: Will users actually want this? - **Usability**: Can users figure out how to use it? - **Feasibility**: Can we build it with our technology? - **Viability**: Can we sustain this business model? For new products, an additional **Go-to-Market** dimension is evaluated. ### 4. Prioritize by Impact and Risk Using the `prioritize-assumptions` skill (lines 56-63), the command maps each assumption on an **Impact × Risk** matrix. This highlights "leap-of-faith" assumptions—those with high impact and high uncertainty that could kill the idea if wrong. The output ranks assumptions by testing priority, ensuring you validate the riskiest elements first. ### 5. Design Validation Experiments The workflow then executes `brainstorm-experiments-existing` or `brainstorm-experiments-new` (lines 68-74), proposing 1-2 experiments per high-priority assumption. Each experiment specifies: - **Method**: A/B test, fake door, pretotype, concierge test, or other validation technique - **Success criteria**: Measurable thresholds for validation - **Effort required**: Time and resource estimates - **Sequencing**: Order of execution ### 6. Generate a Discovery Plan All outputs are assembled into a comprehensive markdown **Discovery Plan** (lines 77-112). According to the source specification, this document includes: - Discovery context and constraints - Brainstormed ideas with selection rationale - Critical assumptions categorized by risk type - Validation experiments with detailed specifications - Timeline and decision framework ### 7. Suggest Next Steps Finally, the command offers follow-up actions (lines 117-123) such as generating a PRD, drafting user interview scripts, setting up success metrics, or creating effort estimates for the engineering team. ## How to Invoke the /discover Command You can initiate the workflow in two ways: **Direct invocation with context:** ```markdown /discover Smart notification system for our project management tool ``` **Interactive mode** (lets the assistant gather context): ```markdown /discover ``` The assistant will then ask: *"Is this an existing product or a new idea?"* ### Typical Interaction Flow ```text Assistant: "What are you exploring? What do you already know? What decisions will this discovery inform?" User: "We need a notification feature for task updates – we already have user feedback showing they miss alerts." Assistant (Brainstorm phase): 1️⃣ PM: "Configurable push notifications" 2️⃣ Designer: "In-app toast with actionable button" 3️⃣ Engineer: "Webhooks to external services" ... User: "Let's keep ideas 1, 2, 5." Assistant (Identify Assumptions): • Value: Users will actually enable notifications • Usability: The toast won't be intrusive • Feasibility: Backend can handle real-time pushes ... ``` ## Key Source Files The `/discover` command orchestrates these specialized skills: - **[`pm-product-discovery/commands/discover.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/commands/discover.md)** – Full command definition and workflow orchestration - **[`pm-product-discovery/skills/brainstorm-ideas-existing/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/brainstorm-ideas-existing/SKILL.md)** – Multi-perspective ideation for existing products - **[`pm-product-discovery/skills/identify-assumptions-existing/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/identify-assumptions-existing/SKILL.md)** – Risk surfacing for current products - **[`pm-product-discovery/skills/prioritize-assumptions/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/prioritize-assumptions/SKILL.md)** – Impact × Risk matrix ranking - **[`pm-product-discovery/skills/brainstorm-experiments-existing/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/brainstorm-experiments-existing/SKILL.md)** – Validation experiment design ## Summary - **The `/discover` command** provides a structured 7-phase workflow for product discovery in the PM Skills framework - **Automatic skill selection** adapts the workflow for existing products versus new ideas - **Seven phases** move from context gathering through brainstorming, assumption identification, prioritization, experiment design, plan creation, and next steps - **15-30 minute sessions** produce a complete markdown Discovery Plan ready for execution - **Checkpoint-based interaction** allows you to steer, skip, or deep-dive at any stage ## Frequently Asked Questions ### What is the difference between existing and new product discovery in /discover? When using the `/discover` command, the workflow automatically selects between "existing" and "new" skill variants based on your initial input. For **existing products**, the command focuses on continuous discovery and feature enhancement, evaluating assumptions across Value, Usability, Feasibility, and Viability. For **new products**, the workflow adds a **Go-to-Market** dimension to assumption identification and places greater emphasis on initial market validation and pretotype experiments. ### How long does a typical /discover session take? According to the implementation in [`pm-product-discovery/commands/discover.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/commands/discover.md), a complete `/discover` session is designed to take **15-30 minutes**. This timeframe accommodates the seven-phase workflow while maintaining momentum through checkpoint-based interactions that let you quickly select ideas, prioritize assumptions, and approve experiment designs without getting bogged down in unnecessary detail. ### Can I skip phases in the /discover workflow? Yes, the command structure includes **checkpoints** at each phase that allow you to steer the conversation, skip steps, or dive deeper where needed. While the workflow is designed to follow the seven-phase sequence for completeness, you can indicate when you already have completed research (skipping context gathering) or when you have predefined assumptions ready for prioritization. ### What output format does the /discover command generate? The command generates a markdown **Discovery Plan** that includes the discovery context, brainstormed ideas with your selections, critical assumptions categorized by risk type (Value, Usability, Feasibility, Viability), validation experiments with methods and success criteria, a timeline, and a decision framework. This document is assembled during phase 6 (lines 77-112) and serves as the primary deliverable for sharing with stakeholders or guiding development.