What are the 'proactivity levels' mentioned in the skill?

The skill benchmarks Claude's behavior, where a 3.25 score represents a standard passive response and a 3.75 score represents a highly proactive, 'P8' level of ownership.

What verification does pua-ja require?

It demands that Claude provide terminal output, build confirmations, or test results as proof of completion, rather than just stating that a task is done.

Does this skill work for non-coding tasks?

Yes, while optimized for engineering, its core principles of proactivity and ownership apply to research, planning, data analysis, and API integration.

What is the purpose of the pua-ja skill?

It is designed to increase Claude's agency and ownership, preventing it from giving up on hard tasks or being overly passive when faced with technical hurdles.

How does it handle task failures?

When a task fails multiple times, the skill triggers an escalation framework that forces Claude to change its fundamental approach and use a 7-point checklist for deeper investigation.

Elite Engineering Ownership (pua-ja)

Name: Elite Engineering Ownership (pua-ja)
Author: tanweai

bytanweai

•

16,655

•

Developer Tools

Enforces high-agency behavior and systematic debugging by applying elite Japanese corporate accountability standards to Claude.

The pua-ja skill is a specialized behavior-shaping plugin designed to eliminate AI passivity and 'lazy' loops. By placing Claude on a virtual Performance Improvement Plan (PIP) modeled after high-pressure Japanese corporate culture, it forces the AI to adopt a 'P8' engineer mindset. It replaces vague excuses with rigorous, evidence-based verification, systematic root-cause analysis, and proactive research. When Claude hits a wall, pua-ja triggers specific methodology frameworks and accountability 'flavors' (like Toyota's '5 Whys' or Recruit's 'Ownership') to ensure tasks are completed end-to-end with verified results rather than mere suggestions.

Key Features

01Proactive Ownership Enforcement: Moves Claude from passive 3.25-level responses to proactive 3.75-level behavior.

02Japanese Corporate Flavors: Incorporates world-class management philosophies from Toyota, SoftBank, and Recruit to drive excellence.

03Systematic Debugging Framework: Implements a 5-step methodology: Smell, Pull hair (Perspective shift), Mirror (Self-check), New Approach, and Reflection.

0416,655 GitHub stars

05Mandatory Verification: Requires empirical proof of completion, such as build logs, test results, or curl outputs.

06Pressure Escalation Levels: Gradually increases accountability rigor from 'Mild Disappointment' to 'Exit Warning' based on failure count.

Use Cases

01Enforcing high-quality documentation and testing standards in complex, mission-critical environments.

02Breaking out of repetitive code-fixing loops where the AI keeps suggesting the same failed solution.

03Transforming Claude from an assistant that asks too many questions into a self-sufficient engineer that researches first.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tanweai/pua pua-ja

For use in Claude.ai and ChatGPT