/discover Command Workflow: How It Chains Product Discovery Skills

The /discover command in the phuryn/pm-skills repository orchestrates a seven-step product discovery workflow that chains four core AI skills—brainstorming, assumption identification, risk prioritization, and experiment design—to transform raw product ideas into actionable validation plans.

The pm-product-discovery plugin within the open-source phuryn/pm-skills repository provides a structured approach to product validation through its /discover command. Defined in pm-product-discovery/commands/discover.md, this command automates a complete discovery cycle by sequentially invoking specialized skills that guide product managers from ideation to testable hypotheses. Understanding this workflow reveals how modular AI prompts can systematically de-risk product decisions.

The Seven-Step /discover Workflow Explained

The /discover command implements a structured pipeline that processes user input through seven distinct phases, with steps 2 through 5 invoking specific skill modules located in the pm-product-discovery/skills/ directory.

Step 1 – Context Capture

The workflow initiates by determining the discovery context. The system queries whether the idea targets an existing product (triggering continuous discovery mode) or a new product (initial discovery mode), while simultaneously ingesting any user-supplied research files to establish baseline context.

Step 2 – Brainstorm Ideas

The command invokes perspective-based brainstorming skills to generate solution candidates. Depending on the context from Step 1, it executes either brainstorm-ideas-existing or brainstorm-ideas-new, defined in pm-product-discovery/skills/brainstorm-ideas-existing/SKILL.md and its new-product counterpart. These skills generate ideas from three distinct viewpoints—Product Manager, Designer, and Engineer—then present the results for user selection.

Step 3 – Identify Assumptions

For each selected idea, the workflow surfaces risky assumptions across four risk domains: Value, Usability, Viability, and Feasibility. New product discovery includes additional GTM and strategy categories. This step leverages the identify-assumptions-existing or identify-assumptions-new skills located in pm-product-discovery/skills/identify-assumptions-existing/SKILL.md.

Step 4 – Prioritize Assumptions

Assumptions are mapped onto an Impact × Risk matrix to identify "leap-of-faith" assumptions—those combining high potential impact with high uncertainty. The prioritize-assumptions skill ( residing in pm-product-discovery/skills/prioritize-assumptions/SKILL.md) ranks these to determine validation priority.

Step 5 – Design Experiments

For top-ranked assumptions, the command chains either brainstorm-experiments-existing or brainstorm-experiments-new to propose validation methods. These skills, defined in pm-product-discovery/skills/brainstorm-experiments-existing/SKILL.md and its new-product variant, suggest 1-2 experiments (such as A/B tests, fake doors, or pretotypes) complete with success criteria, effort estimates, and sequencing.

Step 6 – Assemble Discovery Plan

All generated artifacts—including selected ideas, mapped assumptions, and designed experiments—are compiled into a Markdown discovery plan. This document is formatted for immediate persistence to the user's workspace, creating a record of the complete discovery process.

Step 7 – Offer Next Steps

The command concludes by proposing concrete follow-up actions such as creating a PRD, drafting interview scripts, setting up success metrics, or estimating engineering effort, bridging the gap between discovery and execution.

The Four Core Skills Chained by /discover

The /discover command achieves its modularity by chaining four distinct skill categories that branch based on product maturity:

1. Brainstorming Skills: brainstorm-ideas-existing and brainstorm-ideas-new generate multi-perspective solution ideas from pm-product-discovery/skills/brainstorm-ideas-existing/SKILL.md and its counterpart.
2. Assumption Identification: identify-assumptions-existing and identify-assumptions-new surface risky hypotheses across value, usability, viability, and feasibility domains.
3. Risk Prioritization: prioritize-assumptions maps risks onto an Impact × Risk matrix to identify validation priorities.
4. Experiment Design: brainstorm-experiments-existing and brainstorm-experiments-new design concrete validation tests with defined success metrics.

Each skill operates as a standalone module accessible via individual commands, but the /discover workflow ensures they execute in the optimal sequence for comprehensive validation.

Usage Examples

Run a discovery session on an existing feature:

/discover "Smart notification system for our project-management tool"

Initiate discovery for a new AI product:

/discover "AI writing assistant for non-native speakers"

Interactive workflow after user selects specific ideas:

User: "Here are the top 3 ideas I want to validate."
/discover (continues with assumption identification, prioritization, and experiment design)

Summary

• The /discover command in phuryn/pm-skills implements a seven-step product discovery workflow defined in pm-product-discovery/commands/discover.md.
• It chains four core skill categories: brainstorming, assumption identification, prioritization, and experiment design.
• Skills automatically branch into "existing" and "new" product variants based on initial context capture.
• The workflow generates a Markdown discovery plan ready for immediate workspace integration.
• Each underlying skill can be invoked independently, but /discover guarantees complete end-to-end process execution.

Frequently Asked Questions

What is the difference between the "existing" and "new" product skills?

The "existing" variants (such as brainstorm-ideas-existing and identify-assumptions-existing) focus on continuous discovery for current products with established user bases. The "new" variants include additional GTM and strategy assumption categories critical for zero-to-one product launches, as implemented in their respective SKILL.md files under pm-product-discovery/skills/.

Can I run individual skills without executing the full /discover workflow?

Yes. Each skill functions as a standalone command that can be invoked independently (for example, via /brainstorm or /prioritize-assumptions). The /discover command simply orchestrates these skills in sequence according to the workflow defined in pm-product-discovery/commands/discover.md.

How does the prioritize-assumptions skill determine which assumptions to test first?

The skill plots assumptions on an Impact × Risk matrix, prioritizing "leap-of-faith" assumptions that combine high potential impact with high uncertainty. These represent the riskiest hypotheses that, if invalidated, would most significantly alter the product direction, as detailed in pm-product-discovery/skills/prioritize-assumptions/SKILL.md.

Where are the experiment design skills located in the repository?

The experiment design logic resides in pm-product-discovery/skills/brainstorm-experiments-existing/SKILL.md for continuous discovery and pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md for new product validation. These files define the prompt logic for generating A/B tests, fake doors, pretotypes, and other validation methods.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}