pua

Why we included this project

Most agent skills promise better prompts; this one takes a deliberately provocative route. Rather than just improving prompts, it layers a behavioral system onto coding agents like Claude Code, Codex CLI, and Cursor. When the model starts to give up, the skill escalates through pressure levels, pushing systematic debugging checklists, fact-checking, and proactive follow-through instead of letting the agent spin on the same failed approach. The corporate PUA rhetoric and PIP threats are mostly a joke, but the method underneath is real: three red lines, a five-step debugging process, and fourteen corporate problem-solving flavors. Teams whose agents stall on stubborn bugs or stop after surface fixes will get the most out of it, because those are exactly the lazy patterns the skill targets. It is a prompt and skill artifact, not a standalone tool, so plan to wire it into an existing agent workflow rather than run it on its own.