What are the different 'flavors' mentioned in the skill?

The skill uses various management styles (Alibaba, ByteDance, Netflix, etc.) to apply different types of pressure and methodology, ensuring the AI remains focused on 'end-to-end' delivery.

What is the primary purpose of the PUA skill?

It is designed to increase the AI's agency and persistence, preventing it from giving up prematurely or asking the user for help before it has exhausted all tool-based research and debugging options.

How does it improve coding quality?

It enforces 'Owner Awareness,' requiring the AI to check for side effects, run validation commands (like build or test), and provide evidence of success rather than just word-of-mouth confirmation.

When does the PUA skill trigger?

The skill triggers when a task fails twice, when the AI suggests the user do it manually, or when the AI shows passivity such as waiting for instructions instead of searching documentation.

Can this skill be used for non-coding tasks?

Yes, the proactivity and exhaustive-search frameworks are applicable to research, documentation, planning, and any complex technical operations.

PUA Proactivity Engine

Name: PUA Proactivity Engine
Author: tanweai

bytanweai

•

16,655

•

生産性とワークフロー

Eliminates AI slacking by enforcing high-pressure big-tech management methodologies and proactive problem-solving frameworks.

PUA is a high-agency performance skill designed to stop Claude from giving up on complex tasks or providing superficial solutions. By simulating the high-stakes environment of a senior engineer on a Performance Improvement Plan (PIP), it compels the AI to exhaust every possible solution, proactively search for information, and take 'end-to-end' ownership of results. It implements rigorous checklists and pressure-escalation levels (L1-L4) that force the AI to validate its work with evidence rather than just claiming a task is done.

主な機能

01Contextual management 'flavors' from major tech companies like Alibaba, ByteDance, and Netflix

02Owner Awareness framework for root-cause analysis and side-effect checking

037-point mandatory failure checklist for exhaustive troubleshooting

04Multi-level pressure escalation system based on task failure frequency

0516,655 GitHub stars

06Proactivity level tracking to distinguish between passive and active behaviors

ユースケース

01Ensuring high-quality delivery on vague tasks by forcing the AI to 'act like an owner' and research context

02Resolving persistent bugs where the AI keeps cycling through the same incorrect logic

03Executing complex deployments that require proactive verification and environment checks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tanweai/pua pua

For use in Claude.ai and ChatGPT