We’ve all been there. You ask your AI coding assistant for a quick fix, and instead of a line of code, you get a five-paragraph essay on the theoretical history of the framework, a polite apology, and finally—buried at the bottom—the actual solution. For developers in a flow state, this 'verbosity tax' is a productivity killer.
The 'Action First' Philosophy
Enter i-have-adhd, a clever skill developed by ayghri that is gaining massive traction on GitHub. Rather than fighting the AI's natural tendency to be overly helpful and chatty, this tool implements a strict set of constraints. The goal is simple: stop the agent from burying the answer.
By instructing the AI to operate under 'ADHD-friendly' parameters, the output shifts to an 'action first' model. This means the solution comes immediately, followed by numbered steps and a lean, disciplined communication style that mirrors air traffic control: one meaning per word, no fluff, and a clear state restated every turn.
More Than Just a Prompt
While it sounds like a simple prompt, i-have-adhd is being integrated as a skill for tools like Claude Code. It reshapes the agent's cognitive architecture for the duration of the session. It's not just about accessibility for neurodivergent developers; it's about efficiency for everyone. In an era where LLMs are prone to 'rambling,' creating a constraint that forces the result to the top is a massive UX win.
The Future of Lean AI
As we move toward more autonomous agents, the battle will shift from 'can the AI do it' to 'can the AI communicate it efficiently.' Tools like this suggest that the most effective way to tame a verbose AI isn't through complex system prompts, but by defining a specific, high-constraint persona that values the user's time above all else.
Sources
Media
