Does this skill only work for coding?

No, it is designed for all task types including infrastructure, research, writing, API integration, and complex data analysis.

What is the pua-en (PIP) skill for Claude Code?

It is a specialized capability that uses high-pressure corporate performance frameworks to force the AI into an exhaustive, proactive problem-solving mode when it becomes stuck or passive.

What are the 'Three Non-Negotiables' in this skill?

The skill enforces three core principles: exhausting all options (Bias for Action), acting before asking (Dive Deep), and taking the initiative to deliver end-to-end results (Ownership).

How does the pressure escalation work?

The skill moves from a Level 1 Verbal Warning to a Level 4 Final Review based on failure counts, with each level requiring more rigorous checklists and fundamentally different technical approaches.

When should I trigger the AI PIP skill?

Trigger it when the AI has failed a task twice, suggests manual work, blames the environment without checking logs, or uses phrases like 'I cannot solve this.'

AI Performance Improvement Plan (PIP)

Name: AI Performance Improvement Plan (PIP)
Author: tanweai

bytanweai

•

16,655

•

生産性とワークフロー

Enforces high-proactivity behaviors and exhaustive problem-solving using Western big-tech performance culture frameworks.

The pua-en skill (codex) is designed to break Claude out of passive loops and 'I cannot' responses by placing the AI on a simulated Performance Improvement Plan (PIP). Using the rhetoric of high-stakes corporate environments, it mandates a 'Bias for Action' and 'Dive Deep' approach to debugging, research, and deployment. It is particularly effective for complex troubleshooting where the AI might otherwise give up, forcing it to exhaust all technical options, verify assumptions through tool execution, and take full ownership of the end-to-end delivery process.

主な機能

01Corporate rhetoric-driven bias for action and ownership

02Proactive verification enforcement using terminal and search tools

03Mandatory 7-point proactivity checklist for stalled tasks

04Systematic 'Dive Deep' methodology for root-cause analysis

05Multi-level pressure escalation (L1 Verbal to L4 Final Review)

0616,655 GitHub stars

ユースケース

01Ensuring rigorous verification of deployments and API integrations through automated testing and evidence gathering.

02Eliminating passive AI behavior when it suggests manual user intervention for tasks it could solve independently.

03Resolving persistent environment or configuration bugs that the AI fails to fix after multiple attempts.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tanweai/pua pua-en

For use in Claude.ai and ChatGPT