What Is the Purpose of Command Files in the PM Skills Repository?

Command files in the PM Skills repository serve as metadata-driven specifications that define slash-commands (e.g., /tailor-resume), enabling Claude and compatible AI assistants to execute complex, multi-step product management workflows by chaining together individual skills.

The phuryn/pm-skills repository implements a modular AI assistant marketplace where commands/*.md files function as the primary user interface layer. These markdown files define the purpose and behavior of slash-commands that users invoke to trigger sophisticated workflows. Understanding the purpose of these command files is essential for extending the platform or leveraging its full automation capabilities.

How Command Files Define the Slash-Command Interface

Command files are declarative configuration files stored within plugin directories. Each file in commands/*.md maps a user-facing slash-command to a specific workflow orchestration.

When a user types /command-name, Claude reads the corresponding markdown file at runtime to:

• Generate help text and validate input arguments
• Load the associated skills required for the workflow
• Execute the step-by-step process defined in the file's Workflow section

According to the repository's README, "Commands are user-triggered workflows invoked with /command-name. They chain one or more skills into an end-to-end process" – README § How It Works.

Core Capabilities of Command Files

User-Triggered Workflow Orchestration

Each command file acts as an orchestration layer that transforms static skills into interactive actions. For example, the /tailor-resume command defined in pm-toolkit/commands/tailor-resume.md accepts a resume and job description, extracts keywords, runs the review-resume skill, and returns a rewritten resume with gap analysis.

This mechanism allows single command invocations to trigger complex sequences:

Workflow:
1. Extract JD keywords (skill: extract-keywords)
2. Run resume review (skill: review-resume)
3. Generate tailored version (skill: rewrite-resume)
4. Output comparison report

Self-Documenting Interface Generation

Command files serve as the single source of truth for both execution logic and user documentation. The markdown schema includes:

• Description: Short explanation of the command's purpose
• Arguments: Input hints and validation rules
• Workflow: Step-by-step execution plan

Claude parses these sections to provide real-time assistance without hardcoded help text, making the system self-documenting.

Cross-Plugin Coordination

Commands live inside specific plugins (e.g., pm-toolkit, pm-execution, pm-product-discovery) but can chain skills from other plugins. The /discover command in pm-product-discovery/commands/discover.md executes a four-step discovery flow:

brainstorm-ideas → identify-assumptions → prioritize-assumptions → brainstorm-experiments

This architecture provides a cohesive experience that mirrors real product-management processes while maintaining modular boundaries.

Extensible Architecture

Adding new functionality requires only creating a new markdown file following the established schema. The marketplace loader automatically detects files in */commands/*.md directories, making the system modular and easy to extend without modifying core code.

Practical Command File Examples

The following table illustrates how specific command files map user invocations to skill orchestrations:

Command	User Invocation	File Location	Workflow Description
Tailor Resume	/tailor-resume [upload resume] Here’s the JD: …	pm-toolkit/commands/tailor-resume.md	Accepts resume and JD, extracts keywords, runs review-resume skill, returns rewritten resume with keyword-gap analysis.
Review Resume	/review-resume [upload my-resume.pdf]	pm-toolkit/commands/review-resume.md	Calls review-resume skill to audit against 10 best-practice criteria and returns feedback.
Write PRD	/write-prd New feature: smart notifications	pm-execution/commands/write-prd.md	Chains create-prd skill to produce an 8-section product requirements document.
Discover	/discover AI-powered meeting summarizer	pm-product-discovery/commands/discover.md	Executes four-step discovery flow: brainstorm-ideas → identify-assumptions → prioritize-assumptions → brainstorm-experiments.
Plan Launch	/plan-launch AI-code-review tool	pm-go-to-market/commands/plan-launch.md	Runs gtm-strategy skill suite to produce a complete launch plan from beachhead segment to messaging.

Key File Locations in the Repository

Command files are distributed across plugin directories, each representing a product management domain:

• pm-toolkit/commands/tailor-resume.md – Aligns resumes to job descriptions using the review-resume skill
• pm-toolkit/commands/review-resume.md – Performs comprehensive PM-resume audits
• pm-execution/commands/write-prd.md – Generates full product requirements documents
• pm-product-discovery/commands/discover.md – Executes end-to-end discovery workflows
• pm-go-to-market/commands/plan-launch.md – Builds complete GTM launch plans
• pm-data-analytics/commands/write-query.md – Translates natural-language requests into SQL queries
• pm-ai-shipping/commands/ship-check.md – Produces shipping packets for AI-built code

These files collectively expose the functional surface of the marketplace, allowing users to invoke powerful, multi-step product-management workflows with a single slash command.

Summary

• Command files in commands/*.md are metadata-driven specifications that define slash-command interfaces for AI assistants.
• They orchestrate multiple skills into unified, user-triggered workflows such as resume tailoring or PRD generation.
• Files are self-documenting, automatically generating help text and validation rules at runtime.
• Commands are distributed across plugin directories (e.g., pm-toolkit/commands/, pm-execution/commands/) to maintain modular architecture.
• The system is extensible—adding new commands requires only creating additional markdown files following the established schema.

Frequently Asked Questions

What is the difference between a command and a skill in PM Skills?

Skills are atomic capabilities—individual functions that perform specific tasks like extracting keywords or auditing a resume. Commands are user-facing workflows defined in commands/*.md files that chain one or more skills together into an end-to-end process. For example, the /tailor-resume command orchestrates the review-resume skill along with other steps to provide a complete resume optimization workflow.

How do I add a custom command to the PM Skills repository?

Create a new markdown file in the appropriate plugin's commands/ directory (e.g., my-plugin/commands/my-command.md). The file must follow the established schema including a description, arguments section, and workflow steps. The marketplace loader automatically detects new files at runtime, making the command available immediately without core code modifications.

Where are command files located within the repository structure?

Command files are located in commands/*.md subdirectories within each plugin folder. For example, pm-toolkit/commands/tailor-resume.md and pm-execution/commands/write-prd.md. This distributed structure keeps commands organized by product management domain while maintaining a consistent loading mechanism across the entire phuryn/pm-skills repository.

Can commands in one plugin reference skills from another plugin?

Yes. Commands can chain skills from the same plugin or cross-reference skills from other plugins. For instance, the /discover command in pm-product-discovery coordinates multiple skills to create a cohesive discovery workflow, demonstrating how commands serve as the integration layer that unifies capabilities across the entire PM Skills marketplace.

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