# How the /discover Command Chains Skills for Product Discovery > Learn how the /discover command in phuryn/pm-skills chains four skills to create a complete product discovery plan, from initial context to validated experiments. - Repository: [Pawel Huryn/pm-skills](https://github.com/phuryn/pm-skills) - Tags: how-to-guide - Published: 2026-06-23 --- **The `/discover` command in the `phuryn/pm-skills` repository orchestrates four core skills—brainstorm‑ideas, identify‑assumptions, prioritize‑assumptions, and brainstorm‑experiments—into a sequential workflow that generates a complete discovery plan from initial context to validated experiments.** The `/discover` command serves as the primary entry point for product managers using the `phuryn/pm-skills` framework. It transforms fragmented discovery activities into a structured, reusable pipeline by automatically routing inputs through specialized skills based on product stage. This command demonstrates how modular AI skills can be composed to handle complex, multi-stage product discovery workflows without manual context switching. ## Step-by-Step Skill Orchestration The command implementation in [`pm-product-discovery/commands/discover.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/commands/discover.md) defines a seven-stage pipeline that conditionally routes data between stage-specific skill variants. ### Context Capture and Stage Routing The workflow begins by querying the user to determine whether the product is **existing** or **new**, then collects supporting artifacts such as research documents, PRDs, or interview transcripts. This routing logic sets the `product_stage` variable that determines which skill variants are invoked in subsequent steps. ### Divergent Idea Generation Depending on the stage flag, the command invokes either `brainstorm-ideas-existing` or `brainstorm-ideas-new`. These skills, documented in [`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) and [`pm-product-discovery/skills/brainstorm-ideas-new/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/brainstorm-ideas-new/SKILL.md), generate ten candidate ideas from the perspectives of Product Manager, Designer, and Engineer. The user selects promising candidates to feed into the next stage. ### Assumption Surface and Categorization For each selected idea, the workflow calls `identify-assumptions-existing` or `identify-assumptions-new`. According to the source files [`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) and [`pm-product-discovery/skills/identify-assumptions-new/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/identify-assumptions-new/SKILL.md), these skills extract risky assumptions across **Value**, **Usability**, **Feasibility**, and **Viability** categories. For new products, they additionally surface **Go-to-Market**, **Strategy**, and **Team** risks. ### Impact-Risk Prioritization The aggregated assumptions are passed to the `prioritize-assumptions` skill, implemented in [`pm-product-discovery/skills/prioritize-assumptions/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/prioritize-assumptions/SKILL.md). This skill maps assumptions on an **Impact × Risk** matrix, flags **leap-of-faith** assumptions, and returns a ranked list to guide validation efforts. ### Experiment Design and Sequencing The top-ranked assumptions flow into `brainstorm-experiments-existing` or `brainstorm-experiments-new`. As defined in [`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) and [`pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md), these skills propose concrete validation methods—such as A/B tests, fake doors, landing-page pre-launches, and concierge MVPs—complete with success criteria, effort estimates, and execution sequencing. ### Discovery Plan Assembly All outputs are compiled into a markdown discovery-plan document that records the selected ideas, assumptions, experiment matrix, timeline, and decision framework. The plan is saved to the user’s workspace as a file such as [`discovery-smart-notif.md`](https://github.com/phuryn/pm-skills/blob/main/discovery-smart-notif.md). ### Follow-Up Action Routing Finally, the command presents optional next steps—including generating a PRD, drafting interview scripts, setting up metrics, or creating user-story estimates—allowing the user to extend the workflow without leaving the conversational context. ## Example Invocation and Flow The following demonstrates a typical interaction with the chained skill system: ```markdown # Example invocation /discover Smart notification system for our project management tool ``` ```text # Typical interaction flow User: /discover Smart notification system for our project management tool Bot: 1️⃣ What stage is this product? (existing / new) Bot: 2️⃣ Here are 10 brainstormed ideas … (from brainstorm-ideas-existing) User: I’ll keep ideas 1, 3, 5. Bot: 3️⃣ Identifying assumptions for those ideas … (identify-assumptions-existing) Bot: 4️⃣ Prioritizing assumptions … (prioritize-assumptions) Bot: 5️⃣ Designing experiments … (brainstorm-experiments-existing) Bot: 📄 Discovery plan generated – saved as `discovery-smart-notif.md`. Bot: What next? (create PRD, draft interview script, set up metrics, …) ``` ## Technical Architecture of the Skill Chain The chaining mechanism relies on conditional routing based on the `product_stage` variable. Each skill accepts structured input from the previous stage and returns standardized output for the next, creating a pipeline where context persists across invocations. The command handler manages state transitions between the six primary skills, ensuring that outputs from `brainstorm-ideas` feed into `identify-assumptions`, and that prioritized assumptions automatically trigger the appropriate `brainstorm-experiments` variant. This architecture prevents context loss between discrete discovery activities. ## Summary - The `/discover` command acts as an **orchestrator** for four core skills: brainstorm-ideas, identify-assumptions, prioritize-assumptions, and brainstorm-experiments. - It routes execution to **stage-specific variants** (existing vs. new) based on initial user input captured in [`pm-product-discovery/commands/discover.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/commands/discover.md). - The workflow generates a **markdown discovery plan** documenting ideas, assumptions, and experiments with specific file outputs. - Source implementation files are located in `pm-product-discovery/skills/` with each skill defined in its respective [`SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/SKILL.md) file. ## Frequently Asked Questions ### How does the /discover command choose between existing and new product skills? The command prompts the user to specify the product stage during the **Context Capture** phase. If the user selects "existing", the workflow invokes `brainstorm-ideas-existing`, `identify-assumptions-existing`, and `brainstorm-experiments-existing`. For "new" products, it switches to the corresponding `*-new` skill variants defined in the repository. ### What output format does the /discover command generate? The command produces a **markdown discovery plan** saved to the workspace (e.g., [`discovery-smart-notif.md`](https://github.com/phuryn/pm-skills/blob/main/discovery-smart-notif.md)). This document contains the selected ideas, categorized assumptions, prioritized Impact-Risk matrix, and the experiment design with success criteria and effort estimates. ### Can I skip specific steps in the discovery chain? The current implementation in `phuryn/pm-skills` processes the full chain sequentially to ensure completeness. While you cannot skip core steps within a single `/discover` invocation, you can invoke individual skills directly (e.g., `/skill prioritize-assumptions`) if you need to execute only one specific phase with manual inputs. ### What types of experiments does the brainstorm-experiments skill suggest? The skills suggest validation methods including **A/B tests**, **fake door tests**, **landing-page pre-launches**, **concierge MVPs**, and **usability studies**. Each proposal includes specific success criteria, effort estimates, and sequencing recommendations tailored to the assumption being validated, as specified in [`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) and [`pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md`](https://github.com/phuryn/pm-skills/blob/main/pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md).