Product Discovery and Assumption Testing Skills in the pm-skills Repository

The pm-skills repository provides ten specialized skills for product discovery and assumption testing, orchestrated through five high-level commands that automate the entire discovery workflow from interview synthesis to experiment design.

The phuryn/pm-skills repository offers a comprehensive, LLM-powered framework for product managers conducting discovery work. These skills cover the full spectrum of product discovery and assumption testing, including customer interview analysis, risk identification, and lean experiment design. Each skill is defined in standalone markdown files and invoked through CLI-style commands that chain together into reusable workflows.

Core Discovery Skills

The repository organizes its capabilities into discrete skills located in pm-product-discovery/skills/. Each skill targets a specific phase of the discovery process.

Customer Interview Processing

summarize-interview transforms raw interview transcripts into structured JTBD-focused summaries. Located at pm-product-discovery/skills/summarize-interview/SKILL.md, this skill extracts jobs-to-be-done, pain points, and behavioral patterns from unstructured text.

interview-script generates a complete interview guide before you meet customers. As defined in pm-product-discovery/skills/interview-script/SKILL.md, it produces structured scripts covering warm-up questions, JTBD probes, and wrap-up protocols.

Assumption Identification

identify-assumptions-new surfaces risky assumptions for early-stage concepts across eight risk categories. The definition in pm-product-discovery/skills/identify-assumptions-new/SKILL.md maps risks spanning desirability, viability, feasibility, and usability before launch.

identify-assumptions-existing performs targeted assumption discovery for already-launched products. According to pm-product-discovery/skills/identify-assumptions-existing/SKILL.md, this skill stress-tests new features within live product contexts.

Assumption Prioritization

prioritize-assumptions ranks discovered assumptions using an Impact × Risk matrix. The implementation in pm-product-discovery/skills/prioritize-assumptions/SKILL.md outputs a prioritized table suggesting which assumptions to test first based on potential downside and evidence uncertainty.

Experiment Design

brainstorm-experiments-new designs lean experiments for new product assumptions. Defined in pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md, it generates minimal-effort test plans for pre-launch concepts.

brainstorm-experiments-existing creates concrete validation tests for live products. The skill spec in pm-product-discovery/skills/brainstorm-experiments-existing/SKILL.md focuses on experiments that leverage existing user bases and production infrastructure.

Strategic Planning

prioritize-features ranks backlog candidates by impact, effort, risk, and strategic fit. Located at pm-product-discovery/skills/prioritize-features/SKILL.md, this skill operates after assumptions are validated to guide build decisions.

opportunity-solution-tree visualizes the discovery flow from outcome through opportunities to solutions and experiments. The specification in pm-product-discovery/skills/opportunity-solution-tree/SKILL.md generates OST diagrams for stakeholder communication.

metrics-dashboard defines measurement frameworks for validated learning. As specified in pm-product-discovery/skills/metrics-dashboard/SKILL.md, it structures North Star metrics, input metrics, and health metrics for ongoing monitoring.

Command Orchestration

The pm-product-discovery command set chains individual skills into automated workflows. These commands are defined in pm-product-discovery/commands/*.md and provide one-shot execution of complex discovery sequences.

• /pm-product-discovery:discover – Executes the full discovery cycle: interview processing → assumption identification → prioritization → experiment design.
• /pm-product-discovery:brainstorm – Generates ideas and experiments from cross-functional perspectives (PM, Designer, Engineer).
• /pm-product-discovery:interview – Creates an interview script or summarizes a transcript depending on context.
• /pm-product-discovery:setup-metrics – Builds a metrics dashboard for validated learning post-experimentation.
• /pm-product-discovery:triage-requests – Analyzes and prioritizes inbound feature requests using the assumption-testing framework.

Running a Discovery Workflow

To execute a complete product discovery and assumption testing cycle, invoke the orchestration commands with specific inputs.

Full Discovery Session

/pm-product-discovery:discover

This command prompts for:

1. A customer interview transcript (or generates one using /pm-product-discovery:interview).
2. Assumption identification via identify-assumptions-new (for new products) or identify-assumptions-existing (for live features).
3. Prioritization using prioritize-assumptions to build the Impact × Risk matrix.
4. Experiment design through brainstorm-experiments-new or brainstorm-experiments-existing.
5. Metrics definition with metrics-dashboard to establish success criteria.

Manual Assumption Triage

/pm-product-discovery:triage-requests

When you input a CSV of assumptions, the system executes prioritize-assumptions and returns a markdown table:

Assumption	Impact	Risk	Category	Suggested Experiment
Users will adopt a mobile-first UI	High	Low	Usability	A/B test with 2-week pilot
Enterprise buyers require SSO	High	High	Feasibility	Technical spike with security review

Key Implementation Files

File Path	Purpose
pm-product-discovery/README.md	Module overview and skill catalog
pm-product-discovery/skills/summarize-interview/SKILL.md	Interview synthesis specifications
pm-product-discovery/skills/identify-assumptions-new/SKILL.md	New product risk mapping
pm-product-discovery/skills/identify-assumptions-existing/SKILL.md	Existing feature risk assessment
pm-product-discovery/skills/prioritize-assumptions/SKILL.md	Impact × Risk matrix logic
pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md	Pre-launch experiment design
pm-product-discovery/skills/brainstorm-experiments-existing/SKILL.md	Live product validation tests
pm-product-discovery/skills/prioritize-features/SKILL.md	Backlog ranking algorithm
pm-product-discovery/skills/opportunity-solution-tree/SKILL.md	OST visualization format
pm-product-discovery/skills/metrics-dashboard/SKILL.md	Dashboard specification
pm-product-discovery/skills/interview-script/SKILL.md	Interview guide generation
pm-product-discovery/commands/*.md	CLI command definitions

Summary

• The pm-skills repository contains ten specialized skills for product discovery and assumption testing, each defined in individual SKILL.md files.
• Skills cover the complete discovery lifecycle: interview synthesis, assumption identification (new and existing products), risk prioritization, and experiment design.
• Five high-level commands orchestrate these skills into reusable workflows such as /pm-product-discovery:discover and /pm-product-discovery:triage-requests.
• All skill definitions reside in pm-product-discovery/skills/ while command orchestration logic is located in pm-product-discovery/commands/*.md.

Frequently Asked Questions

What is the difference between identify-assumptions-new and identify-assumptions-existing?

identify-assumptions-new maps risks across eight categories for concepts that have not yet launched, focusing on go-to-market uncertainties and initial value propositions. identify-assumptions-existing targets features being added to live products, stress-testing integration risks, cannibalization concerns, and changes to established user workflows.

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

The skill implements an Impact × Risk matrix that scores each assumption based on the severity of being wrong (Risk) and the business consequence of the assumption (Impact). High-impact, high-risk assumptions surface first, and the skill suggests minimal-effort experiments specifically designed to invalidate those critical assumptions quickly.

Can I use individual skills without running the full discover command?

Yes. Each skill operates independently through its specific invocation pattern. For example, you can run /pm-product-discovery:interview to generate a script, or call summarize-interview directly on a transcript without triggering the complete discovery workflow. The high-level commands are conveniences for common sequences, not restrictions on individual skill usage.

Where are the skill prompts actually defined?

The prompt specifications and logic for each skill are stored in SKILL.md files within their respective directories under pm-product-discovery/skills/. For instance, the assumption prioritization logic resides in pm-product-discovery/skills/prioritize-assumptions/SKILL.md, while the command orchestration definitions are located in pm-product-discovery/commands/*.md.

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