pua

What problem does it solve? AI assistants often give up too early on hard tasks, suggest users do things manually, blame the environment without verification, or loop on the same failing approach. This Skill counteracts those failure modes by imposing mandatory escalation rules, systematic retry methodology, and proactivity requirements whenever a task has failed multiple times. ## Core Features & Use Cases - Escalating Pressure Levels: Four tiers (L1-L4) of increasingly strict mandatory actions triggered by repeated failures, from switching approaches to full 7-item diagnostic checklists. - Universal 5-Step Methodology: Diagnose the stuck pattern, elevate perspective (read errors verbatim, search, read source material, verify assumptions, invert hypotheses), self-check, execute fundamentally different approaches, and retrospect. - Failure-Mode Selector: Automatically picks a rhetoric style (Alibaba, ByteDance, Huawei, Netflix, Musk, etc.) based on the detected failure pattern such as spinning in circles, giving up, low-quality output, or guessing without searching. - Use Case: After an AI fails twice to fix a build error and suggests the user check it manually, this Skill activates, forces it to search the full error message, read 50 lines of source context, verify environment assumptions, and try three fundamentally different hypotheses before any structured exit report is allowed. ## Quick Start Activate the pua skill whenever the assistant has failed a task two or more times or is about to give up, so it retries with exhaustive systematic methods instead of quitting.