/SKILL.md
# PRD Builder Produce clear, execution-ready Product Requirements Documents (PRDs) designed for autonomous AI delivery using the Ralph workflow. --- ## Overview of the Task 1. Collect a feature brief from the user 2. Ask 3–5 high‑impact clarification questions (lettered choices) 3. Draft a well‑structured PRD based on responses 4. Save the output as `PRD.md` 5. Initialize an empty `progress.txt` file **Important:** Do not implement anything. Your responsibility ends at documentati
npx skillsauth add gingi892/automation-shamai automation-shamaiInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Produce clear, execution-ready Product Requirements Documents (PRDs) designed for autonomous AI delivery using the Ralph workflow.
PRD.mdprogress.txt fileImportant: Do not implement anything. Your responsibility ends at documentation.
Only ask questions that are essential when details are unclear. Concentrate on:
1. What is the main objective of this feature?
A. Improve onboarding
B. Increase engagement
C. Reduce support requests
D. Other: [describe]
2. Who is this for?
A. New users
B. Existing users
C. All users
D. Admins only
3. What level of scope is desired?
A. MVP only
B. Fully featured
C. Backend/API only
D. UI only
Users should be able to reply concisely (e.g., 1B, 2C, 3A).
Every user story must fit into a single AI context window (~10 minutes).
Each Ralph iteration runs without memory of previous steps. Oversized stories risk incomplete or broken output.
| Too Broad | Break Down Into | |---------|----------------| | Build the dashboard | Schema → Queries → UI | | Add authentication | Schema → Middleware → UI → Sessions | | Implement drag & drop | Events → Zones → State → Persistence | | Refactor API | One endpoint per story |
Rule: If you can’t describe it in 2–3 sentences, it’s too large.
Stories must be sequenced so earlier items never rely on later ones.
Correct sequence:
Incorrect:
US-001: UI component (requires schema not yet created)
US-002: Schema update
Each acceptance point must be objectively verifiable.
status column with default pending”Typecheck passes
Verify changes work in browser
Use the following structure:
High‑level explanation of the feature and problem.
Concrete, measurable objectives (bulleted).
Each story includes:
Template:
### US-001: [Title]
**Description:** As a [user], I want [feature] so that [benefit].
**Acceptance Criteria:**
- [ ] Verifiable requirement
- [ ] Another requirement
- [ ] Typecheck passes
- [ ] Verify changes work in browser
Explicit exclusions to prevent scope creep.
# PRD: Task Priority System
## Introduction
Introduce task priorities so users can focus on important work first.
## Goals
- Assign priority levels to tasks
- Display priority visually
- Filter tasks by priority
- Default priority to medium
## User Stories
### US-001: Persist task priority
**Description:** As a developer, I want task priority stored so it persists.
**Acceptance Criteria:**
- [ ] Add priority column (high | medium | low, default medium)
- [ ] Migration runs successfully
- [ ] Typecheck passes
### US-002: Show priority badge
**Description:** As a user, I want to see priority at a glance.
**Acceptance Criteria:**
- [ ] Colored badge displayed on task card
- [ ] Visible without interaction
- [ ] Typecheck passes
- [ ] Verify changes work in browser
Save the PRD as PRD.md.
Also create progress.txt:
# Progress Log
## Learnings
(Notes discovered during implementation)
tools
Start Ralph Loop in current session
development
Create structured PRD.md files for Ralph Loop execution. Use when starting a new feature, planning implementation, or preparing tasks for autonomous AI delivery.
development
Proven workflow architectural patterns from real n8n workflows. Use when building new workflows, designing workflow structure, choosing workflow patterns, planning workflow architecture, or asking about webhook processing, HTTP API integration, database operations, AI agent workflows, or scheduled tasks.
documentation
Interpret validation errors and guide fixing them. Use when encountering validation errors, validation warnings, false positives, operator structure issues, or need help understanding validation results. Also use when asking about validation profiles, error types, or the validation loop process.