Agent-Skills.md

Agent Skills: PRD Implementation Planning

Bridges approved PRDs to implementation by mapping user stories to tasks with skill assignments, effort estimates, and progress tracking. Use after PRD approval to create actionable implementation plan.

UncategorizedID: meriley/claude-code-skills/prd-implementation-planning
31

Install this agent skill to your local

pnpm dlx add-skill https://github.com/meriley/claude-code-skills/tree/HEAD/skills/prd-implementation-planning

Skill Files

Browse the full folder contents for prd-implementation-planning.

Download Skill

Loading file tree…

skills/prd-implementation-planning/SKILL.md

Skill Metadata

Name
prd-implementation-planning
Description
Bridges approved PRDs to implementation by mapping user stories to tasks with skill assignments, effort estimates, and progress tracking. Use after PRD approval to create actionable implementation plan.

PRD Implementation Planning

Purpose

Transform approved PRD requirements into actionable implementation tasks with:

  • Skill mappings for each task
  • Dependency ordering and priority
  • Effort estimates
  • Progress tracking framework with date/commit tracking

When NOT to Use This Skill

  • PRD is not yet approved (use prd-reviewing first)
  • Need full technical architecture (use sparc-planning after this)
  • Simple features not requiring planning
  • No PRD exists (use prd-writing first)

Workflow

Step 1: Invoke check-history

Gather project context and understand existing patterns.

Step 2: Read and Analyze PRD

  1. Read the complete PRD document
  2. Extract all user stories (US-1, US-2, etc.)
  3. Identify technical domains involved:
    • Database/Entities (Vendure, TypeORM)
    • API (GraphQL, REST)
    • UI (Admin UI, Mantine)
    • Testing (Playwright, unit tests)
    • Infrastructure (Helm, K8s)
    • Documentation

Step 3: Map User Stories to Tasks

For each user story:

  1. Break into implementation tasks

    • One task per logical unit of work
    • Tasks should be completable in 1-8 hours

  2. Assign primary skill

    • Reference the Skill Mapping in REFERENCE.md
    • Choose the most specific skill available

  3. Estimate effort

    • XS: < 1 hour
    • S: 1-2 hours
    • M: 2-4 hours
    • L: 4-8 hours
    • XL: 8+ hours (should be broken down)

  4. Identify dependencies

    • Which tasks must complete first?
    • Use task numbers (e.g., "Task 1, Task 2")

  5. Set priority

    • P0: Critical path, blocks other work
    • P1: High priority, needed for feature completion
    • P2: Nice to have, can be deferred

Step 4: Generate Skill Requirements Table

Summarize all skills needed:

### Skill Requirements

| Domain   | Skills Required           | Purpose                  |
| -------- | ------------------------- | ------------------------ |
| Database | `vendure-entity-writing`  | Define data models       |
| API      | `vendure-graphql-writing` | Create GraphQL endpoints |
| UI       | `mantine-developing`      | Build components         |
| Testing  | `playwright-writing`      | E2E test coverage        |

Step 5: Generate Implementation Tasks Table

### Implementation Tasks

| #   | Task                       | User Story | Skill                     | Priority | Dependencies | Est. |
| --- | -------------------------- | ---------- | ------------------------- | -------- | ------------ | ---- |
| 1   | Create Order entity        | US-1       | `vendure-entity-writing`  | P0       | None         | M    |
| 2   | Add createOrder mutation   | US-1       | `vendure-graphql-writing` | P0       | 1            | L    |
| 3   | Build order form UI        | US-2       | `mantine-developing`      | P1       | 2            | L    |
| 4   | Write order flow E2E tests | US-1,2     | `playwright-writing`      | P1       | 3            | M    |

Step 6: Initialize Progress Tracker

## Implementation Progress

| #   | Task                       | Status  | Started | Completed | Commit |
| --- | -------------------------- | ------- | ------- | --------- | ------ |
| 1   | Create Order entity        | Pending | -       | -         | -      |
| 2   | Add createOrder mutation   | Pending | -       | -         | -      |
| 3   | Build order form UI        | Pending | -       | -         | -      |
| 4   | Write order flow E2E tests | Pending | -       | -         | -      |

### Progress Summary

- **Total Tasks:** 4
- **Completed:** 0 (0%)
- **In Progress:** 0
- **Blocked:** 0
- **Pending:** 4
- **Last Updated:** YYYY-MM-DD

Step 7: Append to PRD Document

Add the Implementation Plan and Progress sections to the PRD document after the existing content (after Timeline & Milestones if present).

Status Values

| Status | Description | Date Fields | | ------------- | -------------- | --------------------- | | Pending | Not started | Both empty | | In Progress | Active work | Started only | | Done | Completed | Both + commit hash | | Blocked | Cannot proceed | Started, blocker note | | Skipped | Not needed | Note explaining why |

Updating Progress

When a task is completed:

  1. Update status to Done
  2. Fill in Completed date (YYYY-MM-DD)
  3. Add commit hash (short form, e.g., abc1234)
  4. Update Progress Summary counts

Commit Message Pattern: Include [PRD Task N] in commit message for auto-tracking:

feat(orders): add createOrder mutation [PRD Task 2]

Integration with Other Skills

Invokes:

  • check-history - Gather context first

Invoked After:

  • prd-reviewing - PRD must be approved

Works With:

  • sparc-planning - For complex tasks needing architecture
  • safe-commit - Auto-updates progress on task completion

Updates:

  • PRD document with Implementation Plan sections

Example: E-commerce Shipping Feature

Given PRD with:

  • US-1: Calculate shipping rates
  • US-2: Display rate options
  • US-3: Track shipments

Generated Implementation Plan:

Skill Requirements

| Domain | Skills Required | Purpose | | -------- | -------------------------- | ---------------------------------- | | Database | vendure-entity-writing | ShippingRate, Shipment entities | | API | vendure-graphql-writing | Rate calculation, tracking queries | | Plugin | vendure-delivery-plugin | Carrier integration | | UI | vendure-admin-ui-writing | Shipping settings page | | Testing | playwright-writing | Checkout flow tests |

Implementation Tasks

| # | Task | User Story | Skill | Priority | Dependencies | Est. | | --- | ----------------------------- | ---------- | -------------------------- | -------- | ------------ | ---- | | 1 | Create ShippingRate entity | US-1 | vendure-entity-writing | P0 | None | S | | 2 | Create Shipment entity | US-3 | vendure-entity-writing | P0 | None | S | | 3 | Add calculateRates query | US-1 | vendure-graphql-writing | P0 | 1 | M | | 4 | Implement carrier adapter | US-1 | vendure-delivery-plugin | P0 | 3 | L | | 5 | Build rate selector component | US-2 | vendure-admin-ui-writing | P1 | 4 | M | | 6 | Add tracking query | US-3 | vendure-graphql-writing | P1 | 2 | M | | 7 | Write shipping E2E tests | US-1,2 | playwright-writing | P1 | 5 | L |

Anti-Patterns to Avoid

  1. Tasks too large - Break XL tasks into smaller pieces
  2. Missing skills - Every task needs a skill mapping
  3. Circular dependencies - Tasks can't depend on each other
  4. Vague tasks - Be specific about what's being built
  5. Skipping estimates - Every task needs effort estimate
  6. Ignoring priority - P0 tasks should be minimal critical path

Resources

  • REFERENCE.md - Complete skill mapping by domain
  • TEMPLATE.md - Copy-paste templates for sections

For comprehensive specification guidance, use the specification-architect agent.