How the `/discover` Command Runs a Full Product Discovery Cycle

The /discover command orchestrates a structured, seven-stage product discovery workflow by chaining specialized markdown-driven AI skills into an interactive, time-boxed process that moves from initial context to experiment design in approximately 15–30 minutes.

The phuryn/pm-skills repository provides a modular framework for product management automation through slash commands. When you run a full product discovery cycle using /discover <prompt>, the command coordinates multiple discrete skills stored in the repository to transform a single user input into a comprehensive, actionable discovery plan.

Command Architecture and Workflow Structure

The /discover command operates as a workflow orchestrator rather than a monolithic script. Defined in pm-product-discovery/commands/discover.md, the command sequences seven distinct stages, each invoking a specific skill from the pm-product-discovery/skills/ directory. This modular architecture allows individual components—such as brainstorming or assumption identification—to be swapped or customized without altering the core workflow logic.

Each stage functions as an interactive checkpoint, enabling users to redirect, skip, or drill deeper into specific areas while maintaining overall time constraints of roughly 15–30 minutes for the complete cycle.

The Seven Stages of the Product Discovery Cycle

1. Context Gathering

The workflow begins by determining whether the discovery targets an existing product (continuous discovery) or a new product (initial discovery). The command prompts for the discovery question, any known data, and the specific decisions the discovery will inform. This contextual foundation ensures subsequent stages align with actual product needs rather than generic templates.

2. Idea Brainstorming

Depending on the context established, the command invokes either the brainstorm-ideas-existing or brainstorm-ideas-new skill, defined in pm-product-discovery/skills/brainstorm-ideas/SKILL.md. This skill generates ten cross-functional ideas spanning product management, design, and engineering perspectives. The user then selects 3–5 ideas to carry forward into the validation pipeline.

3. Assumption Identification

For each selected idea, the command activates the identify-assumptions-existing or identify-assumptions-new skill from pm-product-discovery/skills/identify-assumptions/SKILL.md. This stage surfaces critical assumptions across five risk categories: Value, Usability, Feasibility, Viability, and (for new products) Go-to-Market. Explicitly mapping these assumptions prevents teams from building on unverified beliefs.

4. Assumption Prioritization

The prioritize-assumptions skill, located in pm-product-discovery/skills/prioritize-assumptions/SKILL.md, maps every identified assumption onto an Impact × Risk matrix. This visualization highlights "leap-of-faith" assumptions—those with high impact but high uncertainty—and ranks them for immediate testing. Low-risk, low-impact assumptions are deprioritized to conserve resources.

5. Experiment Design

Using the brainstorm-experiments-existing or brainstorm-experiments-new skill from pm-product-discovery/skills/brainstorm-experiments/SKILL.md, the command designs 1–2 concrete validation experiments for each top-priority assumption. Experiment types include A/B tests, fake door tests, prototypes, pretotypes, and landing pages, each defined with specific success criteria, effort estimates, and timelines.

6. Discovery Plan Assembly

All outputs compile into a markdown discovery plan that records the initial context, generated ideas, selected concepts, critical assumptions, experiment matrix, and a short-term execution timeline. This document is saved directly to the user’s workspace, creating a persistent record of the discovery process and rationale.

7. Next-Step Suggestions

The command concludes by offering optional follow-up actions, including creating a PRD (Product Requirements Document), drafting interview scripts, setting up success metrics, or estimating effort and user stories. These suggestions bridge the gap between discovery and delivery.

Usage Examples

Invoke the command for new product exploration:

/discover AI writing assistant for non-native speakers

Or for existing product feature discovery:

/discover Smart notification system for our project management tool

A typical interactive flow progresses through checkpoint decisions:

/discover Smart notification system for our project management tool
→ Brainstorm ideas (10 suggestions)
→ User selects 4 ideas
→ Identify assumptions for each idea
→ Prioritize assumptions (high-impact / high-uncertainty)
→ Design experiments (e.g., A/B test notification banner, fake-door landing page)
→ Generate discovery plan markdown
→ Offer next steps (create PRD, draft interview script, etc.)

Modularity and Customization

The workflow’s design intentionally separates concerns into discrete skills. Swapping a skill—such as replacing the default brainstorming module with a custom variant—requires only updating the skill reference in pm-product-discovery/commands/discover.md. This structure ensures the product discovery cycle remains adaptable to specific organizational contexts while maintaining the integrity of the overall process.

Summary

• The /discover command executes a seven-stage workflow defined in pm-product-discovery/commands/discover.md.
• It coordinates five core skills: brainstorming, assumption identification, assumption prioritization, experiment design, and plan assembly.
• The process distinguishes between new and existing products through variant skill implementations (e.g., brainstorm-ideas-new vs. brainstorm-ideas-existing).
• Users interact with checkpoint-based decision points that allow redirection while maintaining a 15–30 minute time box.
• Output includes a markdown discovery plan with experiment designs, success criteria, and actionable next steps.

Frequently Asked Questions

What is the typical duration of a /discover session?

A complete product discovery cycle typically requires 15–30 minutes. The workflow is designed to be time-boxed, with each stage optimized for rapid decision-making. Interactive checkpoints allow users to skip optional depth or extend specific stages if complex assumptions require additional analysis.

Can I skip specific stages in the discovery cycle?

Yes. The command architecture includes interactive checkpoints at each stage transition. Users can redirect the workflow, skip non-critical stages, or dive deeper into specific areas like assumption identification. This flexibility ensures the discovery process adapts to available time and information quality without forcing a rigid linear progression.

How does the command differentiate between new and existing products?

During the initial Context Gathering stage, the command determines the product state and routes to appropriate skill variants. For existing products, it invokes brainstorm-ideas-existing, identify-assumptions-existing, and brainstorm-experiments-existing. For new products, it uses the -new variants, which include additional Go-to-Market risk categories and different brainstorming frameworks suited to zero-to-one exploration.

Where is the /discover command defined in the repository?

The complete workflow definition resides in pm-product-discovery/commands/discover.md. Individual skills referenced by the command are stored in their respective directories under pm-product-discovery/skills/, including brainstorm-ideas/, identify-assumptions/, prioritize-assumptions/, and brainstorm-experiments/. The repository’s README.md provides additional context on command composition and skill chaining.

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