Po Refinement — Development skill for Claude Code
Present open product questions as multiple-choice, record decisions and requirement updates.
How to install Po Refinement
Installs to ~/.claude/commands/dyai2025-astro-noctum-po-refinement.md
mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/DYAI2025/Astro-Noctum/HEAD/.claude/commands/po-refinement.md -o ~/.claude/commands/dyai2025-astro-noctum-po-refinement.md Restart Claude Code, or start a new session, for it to be picked up.
What Po Refinement does
description: Present open product questions as multiple-choice, record decisions and requirement updates allowed-tools: Read, Edit, Write, Grep, Glob
Context
The project uses SDLC scaffold with requirements in `1-objectives/requirements/`, decisions in `2-design/decisions/`, and the objectives index in `1-objectives/CLAUDE.objectives.md`. Tasks with "NEEDS REFINEMENT" in Notes require product decisions before implementation.
Your Task
Collect open product questions from tasks.md
Alternatives in Development
- Projectmem — Open-source coding agent memory 766 ★
- Memory Question — Ask Claude's memory system questions about past work, decisions, and context 332 ★
- Kode Agent SDK — Open runtime for building Claude Code & Manus like Agent Product 320 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Cs Decide
Cold-start decision mode -- present options, record choice to DECISIONS.md, update next action
Elenchus
Turns "what do you think?" into multiple choice, then questions you the way Socrates did until the answer that
Save Session Log
Save a structured record of the current session — what was discussed, what was analyzed/built, where files wer
Extract Application Circuit
CRITICAL REQUIREMENT: You MUST analyze the actual PDF document provided. DO NOT hallucinate, guess, or use pri
Extract Boot Config
CRITICAL REQUIREMENT: You MUST analyze the actual PDF document provided. DO NOT hallucinate, guess, or use pri
Quota
Check Agy Gemini 5h/7d quota or record the user's explicit depletion choice.
Related Agents
Stx Architect
Multi-agent wave Architect persona. Reads requirement-verse.html, raises implementation-gap questions back to
Deep Rl Algorithms Curator
Curator for the deep RL algorithms lane of Reinforcement Learning Brain. Use when maintaining source coverage,
Design Explainer
Explains the current code and design decisions in a form ready to present, and prepares answers to likely foll