Claude for ADHD: The Coding Workflow I Built for My Brain

A Hacker News submission titled **Claude for ADHD: The Coding Workflow I Built for My Brain** points to a write-up on chudi.dev about a personal coding workflow built around **Claude**, with related tooling context that also names **Codex** and **Fable**. At submission it had 2 points and 0 comments, so the piece is still early in public discussion rather than a widely debated thread.

The article frames coding assistance not as a generic chat session but as a workflow shaped for an ADHD attention pattern: externalizing structure, reducing context-switching cost, and keeping the model inside a repeatable loop instead of open-ended freeform prompting. The stack referenced in the source line—**Claude**, **Codex**, and **Fable**—sits in the same product area as agent-style coding helpers and orchestration layers that turn a model into something closer to a session partner than a one-shot Q&A box.

For engineers who already use coding agents, the useful angle is process design rather than model brand. ADHD-friendly constraints—short tasks, explicit next steps, fewer parallel open loops—map cleanly onto how people actually run **Claude** or similar tools day to day: scoped prompts, fixed checklists, and handoff notes that survive an interruption. Builders evaluating agent tooling can treat this less as lifestyle content and more as a case study in interface and habit design around long-running assistant sessions.

That places the post in a crowded market where **Claude**, OpenAI-side coding agents such as **Codex**, and lighter orchestration tools like **Fable** compete on the same job: keep a developer moving through real codebases with less friction. Differentiation is shifting from raw completion quality toward workflow fit—how well a tool holds state, recovers from distraction, and fits an individual’s planning style—which is exactly the ground this article claims.

What to watch next is whether the HN thread and similar posts stay isolated personal notes or feed a clearer pattern: shared templates, ADHD-oriented agent presets, and explicit workflow recipes that travel across **Claude**, **Codex**, and tools like **Fable**. For practitioners, the practical move is to steal the structure—session scaffolding, interruption recovery, single-track tasks—before treating any one vendor as the solution.