poc

Structure proof-of-concept experiments in isolated worktrees with reproducible logs.

Updated Jan 21, 2026
One-click install
npx skills add https://github.com/lixxz/claude-workflow --skill poc
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: poc
Source: https://github.com/lixxz/claude-workflow/tree/main/skills/poc
Command: npx skills add https://github.com/lixxz/claude-workflow --skill poc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Teams struggle to structure uncertain ideas into verifiable experiments. This Skill provides a disciplined framework for building isolated proof-of-concept worktrees, documenting hypotheses, experiments, and learnings to inform decisions.

Core Features & Use Cases

  • Structured worktrees for isolated experiments and clean handoffs
  • Reproducible experiment logs with precise reproduce commands
  • Templates and playbooks for defining hypotheses, success/fail criteria, and evaluation steps
  • Decision gates to decide whether to proceed, pivot, or drop

Quick Start

Initialize a new POC by proposing an idea to create a dedicated worktree and a POC.md to start experiments.

Frequently Asked Questions about poc

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I structure proof-of-concept experiments for reproducible results?

To structure proof-of-concept experiments with reproducible results, use a framework that automates disciplined exploration by guiding hypothesis clarification, enforcing isolated worktrees, and persisting exact commands and logs in a POC.md file.

What is the best way to log experiments and capture exact commands in isolated worktrees?

The best way to log experiments in isolated worktrees is to use a structured logging system that automatically captures exact commands, results, and next steps, ensuring repeatable learnings and clean handoffs across diverse explorations.

How do I define success and fail criteria for a technical hypothesis?

You define success and fail criteria for a technical hypothesis by utilizing provided templates and playbooks, which help structure your evaluation steps and establish clear decision gates to proceed, pivot, or drop the idea.

Can I use isolated git worktrees to manage multiple POC experiments independently?

Yes, you can use isolated git worktrees to manage multiple POC experiments independently, as this approach builds dedicated environments for structured exploration, enabling clean handoffs and preventing conflicts between concurrent hypotheses.

When do I need a decision gate in a proof-of-concept workflow?

You need a decision gate in a proof-of-concept workflow when you must evaluate experiment logs against predefined criteria to decide whether to proceed, pivot, or drop the current hypothesis based on verifiable learnings.

Does this approach require any specific dependencies to start documenting hypotheses?

No specific dependencies are required to start documenting hypotheses, as the framework operates independently to initialize dedicated worktrees and generate POC.md files for structured experiment tracking.