What Skills and Commands Are Available in the pm-product-discovery Plugin?

The pm-product-discovery plugin provides 13 specialized skills and 5 slash commands that automate end-to-end product discovery workflows, from ideation and assumption mapping to customer interview synthesis and feature prioritization.

The pm-product-discovery plugin, maintained in the phuryn/pm-skills repository, structures product management knowledge as discrete, reusable components. Each skill represents a specific domain capability—such as brainstorming experiments or prioritizing assumptions—while commands orchestrate these skills into higher-level workflows accessible via Claude's slash command interface.

Plugin Architecture and File Structure

The plugin organizes its components under the pm-product-discovery directory with a clear separation between interface commands and core logic:

• Plugin manifest: pm-product-discovery/.claude-plugin/plugin.json registers the plugin with Claude and declares available commands.
• Commands: Markdown files in pm-product-discovery/commands/ define slash-command inputs and orchestration logic.
• Skills: Individual directories in pm-product-discovery/skills/ contain SKILL.md files that specify behavior, inputs, and output formats for discrete product management capabilities.

The 13 Product Discovery Skills

The plugin exposes thirteen granular skills, each documented in its own subdirectory under pm-product-discovery/skills/:

Ideation and Experimentation

• brainstorm-ideas-existing: Generates product ideas from PM, Designer, and Engineer perspectives for existing products.
• brainstorm-ideas-new: Produces feature ideas for new products in early discovery phases.
• brainstorm-experiments-existing: Designs experiments to validate assumptions for existing products.
• brainstorm-experiments-new: Creates lean-startup pretotypes for new product validation.

Risk Identification and Prioritization

• identify-assumptions-existing: Maps risky assumptions across Value-Usability-Viability-Feasibility quadrants for existing features.
• identify-assumptions-new: Discovers risks across eight categories including Go-to-Market, Strategy, and Team for new products.
• prioritize-assumptions: Ranks assumptions using an Impact × Risk matrix and suggests corresponding experiments.
• prioritize-features: Evaluates backlog items based on impact, effort, risk, and strategic alignment.
• analyze-feature-requests: Themes and prioritizes incoming feature requests by strategic alignment and effort.

Research and Synthesis

• interview-script: Generates structured customer interview scripts with JTBD probing, warm-up questions, and wrap-up sections.
• summarize-interview: Transforms interview transcripts into structured summaries containing JTBD insights, satisfaction signals, and action items.

Strategic Planning

• opportunity-solution-tree: Constructs Opportunity Solution Trees (OST) that map desired outcomes to opportunities, solutions, and experiments.
• metrics-dashboard: Designs measurement frameworks including North Star metrics, input metrics, health metrics, and alert thresholds.

The 5 Slash Commands

Commands serve as the primary interface for Claude users, mapping slash inputs to one or more underlying skills. The plugin registers five commands in pm-product-discovery/.claude-plugin/plugin.json:

Command	Purpose	Underlying Skills
/pm-product-discovery:brainstorm	Generates ideas or experiments for existing or new products	brainstorm-ideas-existing, brainstorm-ideas-new, brainstorm-experiments-existing, brainstorm-experiments-new
/pm-product-discovery:discover	Executes full discovery cycles: ideation → assumption mapping → experiment design	Combination of brainstorming, identify-assumptions-*, and opportunity-solution-tree
/pm-product-discovery:interview	Creates interview scripts or summarizes transcripts	interview-script, summarize-interview
/pm-product-discovery:setup-metrics	Designs product metrics dashboards	metrics-dashboard
/pm-product-discovery:triage-requests	Analyzes and prioritizes feature request backlogs	analyze-feature-requests

Each command definition resides in pm-product-discovery/commands/ (e.g., discover.md, interview.md).

Command Usage Examples

Invoke commands by passing JSON payloads that specify the operation context. Below are practical examples for each command.

Brainstorm Ideas or Experiments

{
  "command": "/pm-product-discovery:brainstorm",
  "input": {
    "product": "TaskTracker",
    "mode": "existing",
    "focus": "ideas"
  }
}

This returns 5–10 ideas tagged by perspective (PM/Designer/Engineer). Change focus to "experiments" to generate validation experiments instead.

Run Full Discovery Cycles

{
  "command": "/pm-product-discovery:discover",
  "input": {
    "product": "TaskTracker",
    "stage": "full"
  }
}

The command chains multiple skills: it brainstorms ideas, identifies risky assumptions via identify-assumptions-existing, builds an Opportunity Solution Tree, and suggests experiments for high-impact assumptions.

Customer Interview Workflows

{
  "command": "/pm-product-discovery:interview",
  "input": {
    "action": "script",
    "topic": "User onboarding"
  }
}

Set action to "summarise" to trigger summarize-interview and process transcripts into structured insights.

Metrics Dashboard Creation

{
  "command": "/pm-product-discovery:setup-metrics",
  "input": {
    "product": "TaskTracker",
    "northStar": "Monthly Active Users",
    "inputMetrics": ["Sign-ups", "Feature Usage"],
    "healthMetrics": ["Retention", "Error Rate"]
  }
}

Returns a markdown table defining data sources, visualization types, and alert thresholds for each metric.

Feature Request Triage

{
  "command": "/pm-product-discovery:triage-requests",
  "input": {
    "requests": [
      "Add dark mode",
      "Integrate with Slack",
      "Export CSV reports"
    ],
    "criteria": ["impact", "effort", "risk"]
  }
}

Produces a prioritized table with thematic groupings and strategic alignment scores.

Key Implementation Files

Understanding the source structure helps when extending or debugging the pm-product-discovery plugin:

• pm-product-discovery/README.md: Human-readable index of all skills and commands.
• pm-product-discovery/.claude-plugin/plugin.json: Machine-readable manifest declaring command entry points.
• pm-product-discovery/commands/: Contains brainstorm.md, discover.md, interview.md, setup-metrics.md, and triage-requests.md.
• pm-product-discovery/skills/*/SKILL.md: Thirteen skill definition files (e.g., brainstorm-ideas-new/SKILL.md, opportunity-solution-tree/SKILL.md).

Summary

• The pm-product-discovery plugin provides 13 discrete skills covering ideation, assumption mapping, prioritization, interviewing, and metrics design.
• Five slash commands (/pm-product-discovery:brainstorm, /pm-product-discovery:discover, /pm-product-discovery:interview, /pm-product-discovery:setup-metrics, /pm-product-discovery:triage-requests) orchestrate these skills into practical workflows.
• Skills are documented in individual SKILL.md files under pm-product-discovery/skills/, while command orchestration logic lives in pm-product-discovery/commands/.
• The plugin manifest at pm-product-discovery/.claude-plugin/plugin.json registers all capabilities with Claude.

Frequently Asked Questions

How do I install the pm-product-discovery plugin?

The pm-product-discovery plugin is part of the phuryn/pm-skills repository. Clone the repository and place the pm-product-discovery directory in your Claude plugins directory, or point Claude to the .claude-plugin/plugin.json manifest file to load the commands automatically.

Can I use individual skills without invoking the full commands?

Yes. While the five slash commands provide convenient orchestration, you can invoke individual skills directly by referencing their specific SKILL.md definitions in pm-product-discovery/skills/. Each skill accepts defined JSON inputs and produces structured outputs independently of the command layer.

What is the difference between brainstorm-ideas-new and brainstorm-ideas-existing?

The brainstorm-ideas-new skill focuses on early-stage discovery for products not yet built, generating ideas from first principles. The brainstorm-ideas-existing skill assumes a current product context and generates enhancements or adjacent features based on existing user bases and technical constraints. Both map to the /pm-product-discovery:brainstorm command via the mode parameter.

How does the plugin handle sensitive interview data?

The summarize-interview skill processes transcripts locally according to the specifications in pm-product-discovery/skills/summarize-interview/SKILL.md. It structures outputs into JTBD frameworks and action items without transmitting data to external services, though you should verify your specific Claude deployment's data handling policies regarding transcript processing.

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