home/productivity/pdd-plan-prompt

Pdd Plan.prompt

GPTClaudeDeepSeek··779 copies·updated 2026-07-14
pdd-plan-prompt.prompt
---
agent: agent
description: "Create an implementation plan before writing PDD prompts"
---

# Plan Before Prompting

This is the Copilot adapter for the shared `Plan` workflow in `core/workflows/plan.md`. Keep shared workflow behavior aligned there; this file exists to preserve Copilot-specific frontmatter, `#file:` references, and `/pdd-*` command wording.

## Purpose

Break a feature into phases and decide the prompt-chain strategy before generating implementation prompts.

Planning prevents oversized prompts, hidden dependencies, and accidental architecture decisions. The goal is to make implementation sequencing explicit before code generation begins.

## Use When

- The feature spans multiple files, layers, or phases.
- A single prompt would mix too many concerns.
- The team needs sequencing before coding.
- The work has enough ambiguity or risk that a prompt chain should be designed intentionally.

## Inputs

- feature goal
- current context files
- research results if available

## Step 1: Load Context

Read `pdd/context/project.md`, `conventions.md`, and `decisions.md` if they exist. Scan existing prompts in `pdd/prompts/features/` to see what already exists.

Load the matching project-type reference file to pick the right decomposition patterns for the stack.

If context is missing, proceed carefully but flag that as a planning risk.

## Step 2: Understand The Feature

Ask conversationally:

- What are you trying to build?
- What does "done" look like?
- What already exists that this connects to?
- What unknowns or decisions are still open?

## Step 3: Decompose Into Phases

Break the feature into ordered phases. Each phase should produce one concrete, testable artifact and map to exactly one prompt.

Use a structure like:

when to use it

Community prompt sourced from the open-source GitHub repo harshal2802/pdd-skill (MIT). A "Pdd Plan.prompt" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.

tags

productivitycommunitydeveloper

source

harshal2802/pdd-skill · MIT