Claude Code Plugins

Community-maintained marketplace

Feedback
4
0

Comprehensive skill for Product Managers to create PRDs, release plans, user stories, and standalone issues with high-quality acceptance criteria. Supports both full documentation workflows (PRD → Plan → Stories) and quick issue creation (standalone stories, tasks, bugs). All documentation follows design-driven development principles with functional, testable acceptance criteria based on visual designs and user flows.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name product-spec-kit
description Comprehensive skill for Product Managers to create PRDs, release plans, user stories, and standalone issues with high-quality acceptance criteria. Supports both full documentation workflows (PRD → Plan → Stories) and quick issue creation (standalone stories, tasks, bugs). All documentation follows design-driven development principles with functional, testable acceptance criteria based on visual designs and user flows.

Product Spec Kit

A comprehensive skill for Product Managers to create PRDs, release plans, user stories, and standalone issues with high-quality acceptance criteria - all following product management best practices.


Overview

This skill helps Product Managers produce structured, high-quality documentation for product development. It supports both comprehensive documentation workflows (PRD → Plan → Stories) and standalone quick issues (individual stories, tasks, or bugs).

The documentations need to be made using the same language of the user interaction, or the language defined by the user.

Key Capabilities

  1. Create PRDs: Generate complete Product Requirements Documents using workflows/create-prd.md and templates/prd-template.md
  2. Create Plans: Develop release plans with workflows/plan.md and templates/plan-template.md
  3. Create Epics: Define major work themes using workflows/create-epic.md and templates/epic-template.md
  4. Create Issues: Generate user stories, tasks, or bugs using workflows/create-issue.md and templates/stories-template.md
  5. Clarify Documents: Improve quality through targeted questions using workflows/clarify.md
  6. Validate All Documents: Ensure consistency with product principles and templates

File Output Behavior

All generated documents will be saved in a structured specs directory following this organization:

specs/
├── [prd-name]/               # PRD directory (kebab-case)
│   ├── [prd-name].md         # PRD document
│   ├── [plan-name].md        # Release plan document
│   │
│   ├── [epic-name-1]/        # Epic 1 directory (kebab-case)
│   │   ├── [epic-name-1].md  # Epic 1 document
│   │   ├── story-user-login.md
│   │   ├── story-user-logout.md
│   │   └── bug-login-error.md
│   │
│   ├── [epic-name-2]/        # Epic 2 directory
│   │   ├── [epic-name-2].md  # Epic 2 document
│   │   ├── story-admin-dashboard.md
│   │   └── task-setup-database.md
│   │
│   └── [standalone-story].md # Standalone story not part of an epic
│   
└── [another-prd-name]/       # Another PRD directory
    └── ...                   # With its own structure

Naming Conventions

  • PRD Directory: prd-name (kebab-case)
  • PRD File: [prd-name].md (same as directory name)
  • Plan File: [plan-name].md (descriptive name of the plan)
  • Epic Directory: epic-name (kebab-case, descriptive)
  • Epic File: [epic-name].md (same as directory name)
  • Story/Task Files:
    • User stories: story-[short-description].md
    • Bug reports: bug-[short-description].md
    • Technical tasks: task-[short-description].md
    • Research spikes: spike-[short-description].md
  • All files and directories should use kebab-case (lowercase with hyphens)

Fallback Behavior

If the environment doesn't support file system operations, the generated content will be displayed for manual copying and saving in the appropriate directory structure.

Document Hierarchy & Relationships

PRD (Product Requirements Document)

  • Purpose: Defines what to build and why
  • Scope: Product/feature level
  • Detail Level: High-level requirements and goals
  • Template: templates/prd-template.md

Plan

  • Purpose: Organizes work into releases and epics
  • When to create: After PRD approval, before detailed specification
  • Output: Release structure and high-level timeline
  • Template: templates/plan-template.md

Epics

  • Purpose: Groups related user stories that share a common objective
  • When to create:
    • When multiple stories are needed to deliver a feature
    • When work spans multiple sprints
    • When coordinating across teams
  • Template: templates/epic-template.md

Issues (Stories, Tasks, Bugs)

  • Purpose: Defines specific, implementable work items
  • Types:
    • Stories: User-facing functionality
    • Tasks: Technical or non-functional work
    • Bugs: Issues to be fixed
  • Template: templates/stories-template.md

Key Principles

  1. Progressive Elaboration:

    • PRD → Plan → Epics → Stories
    • Each level adds more detail
    • Lower levels must align with higher-level documents
  2. Size & Scope:

    • PRD: Entire product/feature (weeks/months)
    • Epic: Major feature component (2-4 weeks)
    • Story/Task: Small, valuable increment (hours/days)
  3. Value Delivery:

    • Every level should deliver observable value
    • Keep work items small and focused
    • Ensure traceability to business objectives

Core Principles (The Constitution)

All documentation must follow these principles (defined in workflows/product-speckit-constitution.md

Principle I: Complete & Unambiguous Requirements

Every requirement must be clear enough that a reader with no prior context can understand what's needed and why.

Principle II: Design-Driven Acceptance Criteria

Acceptance criteria should be based on visual designs, user flows, and specific UI elements - never written from assumptions.

Principle III: Functional Over Descriptive

Criteria must describe what the system does, not what it looks like. Focus on behavior, actions, and outcomes.

Principle IV: Right Level of Detail

  • PRDs: Macro-level requirements and general rules
  • Stories: Detailed, functional acceptance criteria
  • Each document serves its purpose without overlap

Principle V: Never Assume, Always Validate

When information is missing or ambiguous:

  • Ask specific questions
  • Request designs/mockups
  • Reference previous discussions
  • Mark uncertain items for validation
  • Wait for confirmation before finalizing

Workflows

Full Documentation Workflows

create-prd

Creates a Product Requirements Document following the workflow defined in workflows/create-prd.md using the templates/stories-template.md template.

To create a PRD, simply use the command:

/create-prd

This will guide you through creating a PRD with the appropriate structure and level of detail.

Key Features:

  • Creates a short PRD by default (minimal viable documentation)
  • Option to create a full PRD when explicitly requested
  • Follows the template at templates/prd-template.md
  • Maintains macro-level acceptance criteria (detailed criteria belong in user stories)

Default Short PRD Includes:

  • TL;DR (if sufficient information is available)
  • Context Summary
  • Problem Statement
  • Solution Opportunity
  • High-level Functional Requirements

When to use:

  • Starting a new feature or product
  • Need comprehensive documentation
  • Multiple stakeholders involved
  • Establishing a source of truth for requirements

Usage:

"I want to create a PRD for [feature description]"
"Create a product requirements document for [product]"
"Need a PRD for [initiative]"
"Create a full PRD for [detailed feature]"

Process:

  1. Determines if creating a short (default) or full PRD
  2. Gathers: feature/product description, target users, business goals
  3. Requests designs/mockups if not provided
  4. Validates against constitution principles
  5. Ensures acceptance criteria remain at macro level
  6. Offers download and repository integration options

Note on Micro-interactions:

  • PRDs should not include micro-interactions or UI details
  • These belong in user stories or design documents
  • Focus on business requirements and user value

For complete workflow details, see: workflows/create-prd.md


plan

Creates a structured release plan that serves as a bridge between PRDs and detailed specifications.

Key Outputs:

  • Release structure and timeline
  • High-level epic definitions
  • Dependencies and risks
  • Resource requirements

When to use:

  • After PRD approval
  • Before breaking down into epics and stories
  • When aligning multiple teams

Process:

  1. Review PRD and requirements
  2. Define release strategy
  3. Identify major epics
  4. Map dependencies
  5. Validate with stakeholders

Template: templates/plan-template.md
Workflow: workflows/plan.md

When to use:

  • After PRD is validated
  • Need to break work into releases
  • Planning for execution

Usage:

"Create a plan from my PRD"
"I need a release plan for [feature]"
"Break down my PRD into epics and stories"
"Break down this story/task in subtasks"

Process:

  1. Reads PRD/Epic/Story/Task
  2. Suggests releases and milestones
  3. Breaks into deliveries and epics
  4. Outlines story structure
  5. Validates consistency with PRD/Epic/Story/Task
  6. Asks if user wants to give informations about no priority blocks in default short PRD.

create-epic

Defines major work themes that group related user stories and tasks.

When to use:

  • Breaking down large features into manageable chunks
  • Coordinating work across multiple teams
  • When multiple stories are needed to deliver value

Process:

  1. Define epic scope and objectives
  2. Identify related stories
  3. Set acceptance criteria
  4. Document technical considerations

Template: templates/epic-template.md
Workflow: workflows/create-epic.md


create-issue

Generates detailed user stories, tasks, or bugs with specific acceptance criteria.

When to use:

  • Implementing a specific piece of functionality (Story)
  • Technical or maintenance work (Task)
  • Reporting and fixing issues (Bug)
  • When you need detailed, testable specifications

Key Characteristics:

  • Small, independent units of work
  • Clear acceptance criteria
  • Directly tied to user value or technical need

Usage:

"Create stories from my plan"
"Generate user stories for [epic]"
"I need detailed stories for development"
"Create a bug with these beahviour"

Process:

  1. Reads Plan (and PRD if available)
  2. Requests designs/mockups if not provided
  3. Creates stories for each epic
  4. Bases acceptance criteria on visual elements from designs
  5. Adds technical notes when needed
  6. Validates against constitution

clarify

Asks clarifying questions to improve PRD or Stories/Tasks quality. Use the instructions described on references/clarify.md.

To clarify something, simply use the command:

/clarify [PRD/Story/Task/Bug]

When to use:

  • After initial PRD/Story/Task/Bug creation
  • PRD/Story/Task/Bug feels incomplete or vague
  • Before moving to planning phase

Usage:

"Clarify my PRD"
"Clarify my Story or Epic or Task"
"I have a PRD that needs refinement"
"Ask questions about my PRD"
"Ask questions about my Story"

Process:

  1. Reads existing PRD or Story/Tasks/Bugs (from file or conversation)
  2. Identifies gaps, ambiguities, edge cases
  3. Generates 5-10 targeted questions
  4. Incorporates answers into updated PRD/Story/Task/Bug
To clarify issues

If user asks to clarify stories or tasks, refines stories with clarifying questions. Use the instructions described on references/clarify.md

When to use:

  • Stories feel incomplete
  • Edge cases not covered
  • Before sprint planning

Usage:

"Clarify my stories"
"Ask questions about these stories"
"Refine story [story-id]"

Process:

  1. Reviews stories
  2. Identifies missing criteria or ambiguities
  3. Asks targeted questions about designs and flows
  4. Updates stories with new information

Quick Issue Workflows

create-issue

Create a standalone issue (story, task, or bug) using the workflows/create-issue.md workflow and templates/stories-template.md.

To create an issue, use the command:

/create-issue

This will guide you through creating a well-structured issue with all necessary components, following our standardized format.

When to use:

  • Small, isolated features
  • Design-first approach with existing mockups
  • Clear scope without needing full PRD

Usage:

"Create an issue for [feature description]"
"I need an issue for [description]"
"Create an issue: [what user wants to do]"

Process:

  1. User provides: brief description + designs/mockups
  2. Skill requests designs if not provided
  3. Asks context questions (users, goals, edge cases)
  4. Requests links to related PRDs/docs if available
  5. Generates complete issue with:
    • Issue format
    • Design-based acceptance criteria
    • Technical notes
    • Definition of Done
  6. Validates against constitution

Output: Single issue ready for development


quick-story

Creates a standalone user story without full PRD/Plan workflow.

When to use:

  • Small, isolated features
  • Design-first approach with existing mockups
  • Clear scope without needing full PRD

Usage:

"Create a quick story for [feature]"
"I need a user story for [description]"
"Quick story: [what user wants to do]"

Process:

  1. User provides: brief description + designs/mockups
  2. Skill requests designs if not provided
  3. Asks context questions (users, goals, edge cases)
  4. Requests links to related PRDs/docs if available
  5. Generates complete story with:
    • User story format
    • Design-based acceptance criteria
    • Technical notes
    • Definition of Done
  6. Validates against constitution

Output: Single story ready for development


7. quick-task

Creates a standalone technical task.

When to use:

  • Technical improvements
  • Refactoring needs
  • Infrastructure work
  • Non-user-facing work

Usage:

"Create a task for [technical work]"
"Quick task: [description]"
"I need a tech task for [improvement]"

Process:

  1. User provides: technical description
  2. Asks context questions (why, impact, dependencies)
  3. Requests related docs/ADRs if available
  4. Generates complete task with:
    • Clear objective
    • Technical acceptance criteria
    • Dependencies
    • Testing requirements
  5. Validates against constitution

Output: Single task ready for execution


8. quick-bug

Creates a standalone bug report with reproduction steps.

When to use:

  • Bugs found in production or testing
  • Quick bug documentation needed
  • Urgent fixes

Usage:

"Create a bug report for [issue]"
"Quick bug: [description]"
"Document this bug: [problem]"

Process:

  1. User provides: bug description
  2. Asks for: steps to reproduce, expected vs actual behavior
  3. Requests screenshots/recordings if available
  4. Generates complete bug with:
    • Clear description
    • Reproduction steps
    • Expected behavior
    • Actual behavior
    • Acceptance criteria for fix (follow the same story acceptance criteria format)
  5. Validates against constitution

Output: Single bug ready for triage/fixing


Design-Driven Development

Why Designs Matter

Acceptance criteria should be based on what you can see in designs, not assumptions. This skill actively requests designs and uses them as the source of truth for criteria.

Always verify if the user have some MCP installed to connect to Figma.

When Skill Requests Designs

  • Creating stories (create-stories)
  • Creating quick stories (quick-story)
  • Clarifying ambiguous requirements
  • Validating user flows
  • When criteria feel vague

What to Provide

Best:

  • Figma/Sketch links
  • Interactive prototypes
  • Annotated wireframes
  • User flow diagrams

Good:

  • Screenshots of mockups
  • Wireframes
  • Sketches with notes

Acceptable:

  • Detailed written descriptions of UI
  • References to existing similar features

How Skill Uses Designs

  1. Identifies UI elements: buttons, forms, modals, menus
  2. Maps user interactions: clicks, inputs, navigation
  3. Extracts user flows: entry points, steps, outcomes
  4. Writes functional criteria: based on visible elements and behaviors
  5. Asks clarifying questions: for ambiguous visual elements

Acceptance Criteria Format

Remember the format of the criteria format for PRD and Stories/Tasks when use templates/prd-template.md, templates/quick-issue-template.md and templates/stories-template.md

Document-Specific Acceptance Criteria

For PRDs (Macro-Level)

Focus on the "what" and "why":

  • Define overall product capabilities
  • Set business rules and constraints
  • Establish success metrics

For Epics (Mid-Level)

Bridge between strategy and implementation:

  • Define feature-level requirements
  • Set boundaries for related stories
  • Include integration points
  • Define non-functional requirements
#### Must Have (P0)
1. **User Authentication**
   - Description: Users must be able to securely log in
   - User Value: Protects user data and enables personalization
   - Acceptance Criteria:
     - System must support email/password authentication
     - System must validate email format
     - System must enforce password complexity rules
     - System must provide password reset functionality

For Stories (Detailed, Functional)

Focus on specific, testable behaviors based on designs:

## Acceptance Criteria

### Authentication Flow
- When user clicks "Login" button on homepage, system displays login modal
- When user enters email in email field, system validates format in real-time
- When user enters invalid email format, system displays "Please enter a valid email" error below field
- When user enters password and clicks "Submit", system attempts authentication
- When authentication succeeds, system closes modal and redirects to dashboard
- When authentication fails, system displays "Invalid credentials" error above form
- When user clicks "Forgot Password" link, system displays password reset modal

### Password Reset Flow
- When user enters email in reset form and clicks "Send Reset Link", system sends reset email
- When system sends email, system displays "Check your email for reset link" confirmation
- When user doesn't receive email within 1 minute, system shows "Resend Email" button

Note the difference:

  • PRD/Epics: "System must support authentication" (what, why)
  • Stories: "When user clicks 'Login' button..." (specific, testable, design-based)

Progressive Detailing:

  1. PRD (Why): "System must support user authentication to protect user data"
  2. Epic (What): "Implement authentication system with email/password and social login"
  3. Story (How): "As a user, I can log in with my email and password so I can access my account"
  4. Acceptance Criteria:
    When I enter a valid email and password and click 'Login',
    Then I should be redirected to my dashboard
    And see a success message
    

Key Differences:

  • PRD: Business objectives and high-level requirements
  • Epic: Feature-level scope and technical approach
  • Story/Task: Specific implementation details and UI behavior

Best Practices

Starting a New Feature

  1. Gather designs first - request mockups/wireframes before writing
  2. Create PRD - establish the big picture
  3. Clarify PRD - refine through questions
  4. Create Plan - break into releases and epics
  5. Create Stories - detail with design-based criteria
  6. Clarify Stories - refine before development

Working with Quick Issues

When to use Quick Issue:

  • Small, isolated changes
  • Clear designs available
  • No need for extensive planning
  • Urgent items

Best for:

  • UI improvements
  • Bug fixes
  • Small features
  • Research spikes
  • Single user stories
  • Urgent items

Working with Designs

  1. Always request design files if not provided
  2. Analyze user flows from visual elements
  3. Ask about ambiguous UI elements
  4. Map user actions to acceptance criteria
  5. Validate interpretation with user

Avoiding Assumptions

  1. Ask specific questions when context is missing
  2. Reference previous discussions when available
  3. Mark uncertain items for validation
  4. Wait for confirmation before finalizing
  5. Re-validate after feedback

Troubleshooting

"My criteria feel too vague"

→ Request designs/mockups
→ Ask about specific user flows
→ Use clarify-stories workflow
→ Review constitution Principle II

"Stories don't match designs"

→ Share design files with skill
→ Point out specific elements in designs
→ Ask for criteria revision
→ Validate step-by-step

"Quick issues lack context"

→ Provide more background upfront
→ Share related docs/PRDs
→ Upload design files
→ Reference previous discussions

"Criteria are too detailed for PRD"

→ PRDs should have macro-level requirements
→ Move detailed criteria to stories
→ Keep PRD focused on general rules

"Criteria are too high-level for stories"

→ Stories need detailed, functional bullet points
→ Request designs if not provided
→ Ask for specific user flow clarification
→ Use clarify-stories to refine


Tips for Success

  1. Provide designs early: Share mockups, wireframes, images from the start
  2. Reference previous work: Link to related PRDs, epics, stories
  3. Be specific in questions: Better to over-explain than assume
  4. Validate often: Check criteria match expectations before moving forward
  5. Differentiate levels: PRD = macro, Stories = detailed
  6. Use designs as source: Base criteria on visual elements when available

File Structure

See docs/STRUCTURE.md for complete file structure details.


Getting Started

First time users:

  1. Read the docs/QUICKSTART.md guide
  2. Review rules/product-speckit-constitution.md principles
  3. Try creating a quick story using designs provided by user
  4. Progress to full PRD when ready

Need help?

  • Check docs/INDEX.md for complete documentation reference
  • Review docs/STRUCTURE.md to understand file organization