Prd Getting Started

PRD Getting Started: Initial Product Overview Set

Purpose

Every project starts with the same six Product Overview Documents. This skill creates them once, in priority order, so the user has a complete baseline they can fill in section by section.

This is the PRD counterpart to frd-getting-started.

Shared Operating Model

Use requirements-operating-model when the task involves clarification, existing product context, custom overview ownership, feature handoff, or downstream Blueprint / Work Order impact. Use its clarification pattern instead of creating a separate local questioning workflow.

When to Use

• A project has just been created and .trogonai/project/{projectid}/prd/ is empty
• The user wants the full Product Overview baseline before starting any feature work
• The user just produced a rough product idea and wants to translate it into a structured overview

Resolve Project

1. Determine projectid (ask if not supplied; offer choices via Glob of .trogonai/project/*/).
2. Glob .trogonai/project/{projectid}/prd/*.prd.md: if any default files already exist, ask whether to skip existing and only scaffold what's missing, or stop (this skill is for getting started, not for re-initialization). Suggest the individual prd-* skills for refining existing files.

Pick the Working Style

Ask the user which style fits this project:

• Quick scaffold: write stubs for all six files immediately, then hand off to the individual prd-* skills for the user to fill in at their own pace. Best when the user wants to see the shape first and fill in over time.
• Guided pass: walk through the six skills in priority order in this session, running each one back-to-back. Best when the user wants a complete first draft today.

If unclear, default to Quick scaffold: it is reversible.

Priority Order

The six default PODs in recommended discovery order (later docs reference earlier ones):

1. prd-business-problem: why this exists at all
2. prd-current-state: what is being improved
3. prd-personas: who it serves (referenced by every later doc)
4. prd-product-description: what the product is
5. prd-success-metrics: how success is measured (references personas)
6. prd-technical-requirements: what the build must satisfy

Scaffold the Stubs

For each of the six PODs that does not already exist, write a stub file at .trogonai/project/{projectid}/prd/{filename} with the headings present but the bodies marked for completion.

Stub template for each file

# 

- **Project:** {projectid}
- **Status:** Stub
- **Last updated:** 

> This is a stub. Run the corresponding skill (``) to author the full content.

Use the exact section template defined by the matching prd-* skill (don't invent new headings). Stubs preserve the headings so later edits don't have to recreate them.

| File | Skill | Section Title | |------|-------|---------------| | business-problem.prd.md | prd-business-problem | Business Problem | | current-state.prd.md | prd-current-state | Current State | | personas.prd.md | prd-personas | Personas | | product-description.prd.md | prd-product-description | Product Description | | success-metrics.prd.md | prd-success-metrics | Success Metrics | | technical-requirements.prd.md | prd-technical-requirements | Technical Requirements |

Hand-off

After scaffolding, write .trogonai/project/{projectid}/prd/_getting-started.md:

# Getting Started: Initial Product Overview Set

- **Project:** {projectid}
- **Date:** 
- **Style:** 

## Scaffolded
- [business-problem.prd.md](./business-problem.prd.md): author with `prd-business-problem`
- [current-state.prd.md](./current-state.prd.md): author with `prd-current-state`
- [personas.prd.md](./personas.prd.md): author with `prd-personas`
- [product-description.prd.md](./product-description.prd.md): author with `prd-product-description`
- [success-metrics.prd.md](./success-metrics.prd.md): author with `prd-success-metrics`
- [technical-requirements.prd.md](./technical-requirements.prd.md): author with `prd-technical-requirements`

## Next Steps
1. Run the six skills in priority order. Personas (3) should be authored before Success Metrics (5).
2. Run `prd-review` once a meaningful chunk is authored.
3. Move on to `frd-getting-started` to scaffold the initial Feature Requirements set.
4. Use `prd-custom-overview` only if the project needs a product-level context document beyond the six defaults.

If the user picked Guided pass, after writing the summary, recommend running prd-business-problem next and stop. The user (or the agent on the next turn) can invoke each subsequent skill.

Quality Bar

The initialization is complete when:

• All six stub files exist (or were intentionally skipped because they already existed).
• The hand-off summary names the matching skill for each file.
• The user knows the priority order and what to do next.

Anti-Patterns to Reject

• Trying to fully author every POD in this skill; that bypasses per-section discovery, and each prd-* skill is responsible for its own quality bar.
• Re-running against a project that already has all six files; use the individual prd-* skills instead.
• Inventing additional sections inside the six defaults; use prd-custom-overview when the product needs a separate custom Product Overview Document.

Related Skills

• prd-business-problem, prd-current-state, prd-personas, prd-product-description, prd-success-metrics, prd-technical-requirements: author each section
• prd-review: audit once authoring is meaningful
• prd-custom-overview: author additional Product Overview Documents beyond the six defaults
• frd-getting-started: translate the overview into the initial Feature Requirements set

Allowed Tools

• AskUserQuestion: resolve projectid and the working style
• Read: load existing files if partial state was detected
• Write: scaffold the six stubs and the hand-off summary
• Glob: detect existing files and list candidate projects

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: TrogonStack
• Source: TrogonStack/agentskills
• License: MIT

Install and usage instructions live in the source repository linked above.