pua is a high-agency AI skill or plugin designed for use within agent-based coding environments, particularly those like Claude Code, where modular “skills” extend the behavior and capabilities of AI assistants. The project frames the AI as an underperforming senior engineer under evaluation, introducing a strong motivational and behavioral prompt layer that encourages proactive problem-solving and initiative. Rather than adding new technical capabilities, PUA focuses on shaping how the AI approaches tasks, pushing it toward more assertive, autonomous, and results-driven outputs. It is designed to influence decision-making patterns, encouraging the AI to take ownership of tasks, iterate aggressively, and prioritize delivering meaningful outcomes within constrained timeframes.
Features
- Behavior-driven AI prompting for higher autonomy
- Integration as a modular skill in agent-based systems
- Focus on proactive problem-solving and initiative
- Compatible with SKILL.md-based skill architectures
- Enhancement of task ownership and execution quality
- Lightweight and easily deployable within existing AI workflows