How to Chain Multiple PM Skills for Custom Commands

You chain multiple pm-skills together by creating a Markdown command file in a plugin's commands/ directory that references individual skills from the skills/ folder using the **Apply the <skill-name> skill** pattern, allowing Claude to execute them sequentially while passing data through conversation context.

The pm-skills marketplace by phuryn/pm-skills treats complex product management workflows as composable units. Chaining multiple pm-skills together allows you to build custom commands that automate multi-step processes—from initial discovery to go-to-market launch—using only Markdown files without modifying source code.

Understanding the PM Skills Architecture

The repository organizes functionality into two distinct layers:

• Skills are reusable frameworks defined in pm-<plugin>/skills/<skill-name>/SKILL.md. Each skill contains the logic for a specific product management task, such as brainstorm-ideas-new or prioritize-assumptions.
• Commands are orchestration files stored in pm-<plugin>/commands/<command-name>.md. These Markdown files chain multiple skills together into executable workflows.

When Claude processes a command, it automatically loads and executes referenced skills in sequence, passing intermediate results through the conversation context.

Creating a Custom Command That Chains Skills

To build your own chained workflow, create a new Markdown file in the appropriate plugin directory:

1. Select a plugin folder (e.g., pm-product-discovery or pm-go-to-market).
2. Create a Markdown file in the plugin's commands/ directory (e.g., pm-go-to-market/commands/quick-launch.md).
3. Add YAML frontmatter at the top describing the command and optional arguments.
4. Reference skills in the workflow section using the **Apply the <skill-name> skill** syntax.
5. Save the file—the command becomes instantly available to Claude-Code or Claude-Cowork.

Example: The /discover Command

The /discover command in pm-product-discovery/commands/discover.md demonstrates chaining four distinct skills:

---
description: Run a full discovery cycle
---

# Discovery Command

Apply the **brainstorm-ideas-new** skill.

Then apply the **identify-assumptions-new** skill.

Apply the **prioritize-assumptions** skill.

Finally, apply the **brainstorm-experiments-new** skill.

This sequence executes skills in order: generating ideas, identifying assumptions, prioritizing them, and designing experiments. The skills are defined in pm-product-discovery/skills/brainstorm-ideas-new/SKILL.md, pm-product-discovery/skills/identify-assumptions-new/SKILL.md, pm-product-discovery/skills/prioritize-assumptions/SKILL.md, and pm-product-discovery/skills/brainstorm-experiments-new/SKILL.md.

Building a Complex Workflow: The /quick-launch Example

For a sophisticated go-to-market workflow, create pm-go-to-market/commands/quick-launch.md with the following content:

---
description: End‑to‑end launch workflow that builds an ICP, a GTM plan, and a growth loop
argument-hint: "<product name>"
---

# /quick-launch – Fast Go‑to‑Market Launch

## Workflow

### Step 1: Build the Ideal Customer Profile

Apply the **ideal-customer-profile** skill:

- Ask the user for the product name and primary market.
- Generate an ICP with demographics, behaviors, and needs.

### Step 2: Draft the GTM Strategy

Apply the **gtm-strategy** skill:

- Use the ICP from Step 1.
- Produce a channel mix, messaging pillars, and success metrics.

### Step 3: Design a Growth Loop

Apply the **growth-loops** skill:

- Based on the GTM plan, suggest a self‑reinforcing loop.
- Include high‑level tactics and effort estimates.

When a user runs /quick-launch Super‑AI Writing Assistant, Claude executes ideal-customer-profile (defined in pm-go-to-market/skills/ideal-customer-profile/SKILL.md), passes that output to gtm-strategy (from pm-go-to-market/skills/gtm-strategy/SKILL.md), then feeds both into growth-loops (from pm-go-to-market/skills/growth-loops/SKILL.md).

How Data Flows Between Chained Skills

Data hand-off occurs automatically via conversation context. Each skill's output becomes the next skill's input without explicit variable passing. When you chain pm-skills together, Claude maintains the cumulative state across the entire workflow, allowing downstream skills to reference earlier outputs naturally.

Testing Your Custom Command

To verify your chained command works:

1. Place your Markdown file in pm-<plugin>/commands/.
2. Load the plugin in Claude-Code or Claude-Cowork.
3. Invoke the command: /quick-launch ProductName.
4. Observe the sequential execution of each skill section.

Summary

• Chain multiple pm-skills by creating Markdown files in a plugin's commands/ directory.
• Reference skills using the **Apply the <skill-name> skill** pattern.
• Data flows automatically between skills via conversation context.
• No code changes are required—only Markdown files with YAML frontmatter.
• Example implementations include /discover in pm-product-discovery/commands/discover.md and /quick-launch in pm-go-to-market/commands/quick-launch.md.

Frequently Asked Questions

Do I need to write code to chain pm-skills together?

No. The pm-skills marketplace uses pure Markdown files. You create a command by adding a new .md file with YAML frontmatter in the commands/ directory. The system parses these files to load and execute referenced skills automatically without requiring code compilation or deployment.

Where should I save custom command files?

Save custom commands in pm-<plugin>/commands/<command-name>.md within your plugin folder. For example, pm-go-to-market/commands/quick-launch.md makes the /quick-launch command available immediately upon saving the file.

How does data pass between skills in a chain?

Data passes implicitly through conversation context. Each skill appends its output to the conversation, and subsequent skills in the chain can access this cumulative context. This allows seamless hand-offs where one skill's deliverables become the next skill's inputs without explicit configuration or variable declaration.

Can I chain skills from different plugins?

Yes, though skills typically reside within their parent plugin's skills/ directory. You can reference any skill by name if Claude has loaded the plugin. For explicit loading, use the /plugin-name:skill-name syntax before running your custom command.

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