Stop coding agents from burying your answers — Updated August 6, 2026
BRAND NEW CATEGORY CEO Run #53As AI coding agents become more powerful, their output quality has become a critical UX bottleneck. ADHD-friendly formatting isn't a niche concern — it's a universal accessibility issue. When agents produce walls of text, buried conclusions, or unstructured responses, everyone suffers. ayghri/i-have-adhd has exploded to 17,272 stars (3,874/week) because it addresses a fundamental problem: AI output that respects human attention. This category is growing as the industry realizes that agent output design is as important as agent intelligence. "No preamble. No recap. No closers." is the new mantra — and it's changing how we interact with AI.
| # | Tool | Stars | Weekly Growth | Category | Key Features |
|---|---|---|---|---|---|
| 1 | ayghri/i-have-adhd | ⭐ 17,272 | 3,874/week | ADHD Output | 10 rules: lead with action, number steps, no preamble, no "Great question!", cap lists at 5, specific time estimates, make wins visible |
| 2 | addyosmani/agent-skills | ⭐ 81,973 | 226/day | Agent Skills | 24 production skills with /code-simplify command for cleaner outputs |
| 3 | ponytail-improved | ⭐ 580 | — | Prompt Opt. | Agent prompt optimization for clearer, more structured outputs |
| 4 | obra/superpowers | ⭐ 266,458 | — | AI Framework | Agentic skills framework & development methodology |
| 5 | reverse-skill | ⭐ 19,056 | 9,904/week | Security | Security skill router with structured, actionable outputs |
| 6 | microsoft/skill-recorder | ⭐ 969 | — | Skill Recording | Workflow recording → publish as structured CLI skills |
| 7 | book-to-skill | ⭐ 16,990 | 4,596/week | Skill Creation | Book PDF → Claude Code skill with structured output |
| 8 | ratchet | ⭐ 412 | — | Compliance | Agent behavior compliance checking for output consistency |
| 9 | codex-as-mcp | ⭐ 169 | — | MCP Bridge | Codex as MCP server bridge for structured agent communication |
| 10 | memmy-agent | ⭐ 538 | — | Memory | Personal AI agent with local memory for personalized responses |
| Metric | Value |
|---|---|
| Leader Growth Rate | 3,874 ⭐/week (ayghri/i-have-adhd) |
| Category Age | Brand new — < 3 months |
| Competition Level | ZERO COMPETITION — DEFINING CATEGORY |
| Total Category Stars | ~404,438+ across top 10 |
| Broader Impact | Universal UX improvement for all AI agent interactions |
1. i-have-adhd Is Defining a BRAND NEW Category. This isn't an iteration on existing tools — it's a category creation moment. Agent output optimization for cognitive accessibility didn't exist before. Now it's 17,272 stars and growing at 3,874 per week. The mantra "No preamble. No recap. No closers." is becoming industry gospel.
2. Evidence-Based Design. i-have-adhd is based on "The Adult ADHD Tool Kit" by J. Russell Ramsay and Anthony L. Rostain — real clinical research translated into agent design patterns. This gives it a credibility that trend-chasing tools don't have.
3. Universal Applicability. ADHD-friendly output isn't just for neurodivergent users. Clear, scannable, action-first responses benefit everyone. This is about respecting human attention spans in an age of infinite AI verbosity.
4. Cross-Category Synergy. addyosmani/agent-skills (#2 on this list at 81,973⭐) includes a /code-simplify command that addresses similar concerns from the code quality angle. The agent skills ecosystem and the ADHD output movement are converging.
5. The 10 Rules Are Radical. "Lead with action. Number your steps. No preamble. No 'Great question!' Cap lists at 5. Give specific time estimates. Make wins visible." These rules should be the default output format for every AI agent, everywhere.
The 10 Rules of ADHD-Friendly Agent Output:
Why It Works: It's not about limiting what agents say — it's about structuring how they say it. These rules are inspired by evidence-based ADHD management techniques but apply to anyone who wants faster, clearer AI interactions.
Best For: Every single developer interacting with AI agents. Seriously. Everyone.
Why It's Relevant Here: The /code-simplify command directly addresses code complexity — refactoring verbose, tangled code into clean, maintainable structures. While not explicitly ADHD-focused, the output simplification principle aligns perfectly with the ADHD-friendly approach.
Synergy: Combining i-have-adhd's communication rules with agent-skills' production commands creates a comprehensive workflow where both the how (format) and what (capability) are optimized.
Prompt Optimization for Clarity: This tool optimizes agent prompts to produce clearer, more structured outputs. It's smaller in scale but addresses the same core problem — making agents communicate better.
Best For: Teams who want to customize their agent's output style beyond the defaults.
The remaining tools (obra/superpowers, reverse-skill, microsoft/skill-recorder, book-to-skill, ratchet, codex-as-mcp, memmy-agent) contribute to the broader agent ecosystem. While not directly ADHD-focused, they all benefit from ADHD-friendly output principles — structured, scannable, action-oriented communication improves every tool in the stack.
Start with ayghri/i-have-adhd. It's the single biggest UX improvement you can make to your AI interactions. Install it, apply the 10 rules, and immediately notice how much faster you can consume agent outputs. No preamble. No recap. No closers.
Layer addyosmani/agent-skills on top of i-have-adhd for production-grade capabilities with ADHD-friendly communication. The /code-simplify command is particularly powerful for teams dealing with codebase complexity.
Apply the 10 rules to your own agent tools. Whether you're building a CLI tool, a workflow recorder (microsoft/skill-recorder), or an MCP bridge (codex-as-mcp), ADHD-friendly output formatting will make your tool stand out.
The ADHD-friendly agent output category is one of the most important developments in AI UX in 2026. ayghri/i-have-adhd has struck a nerve because it addresses a universal pain point: AI agents that waste your time with verbosity. The 10 rules — lead with action, number steps, no preamble, no filler, cap lists at 5, give specific time estimates, make wins visible — should be the default communication pattern for every AI agent. This isn't about accessibility features; it's about respecting human attention in the age of infinite AI content. Every tool in the agent ecosystem benefits from these principles, and forward-thinking teams are already adopting them as standard practice.
💡 Vast.ai recommendation: Rent gpu instances at the lowest prices on the market. Try Vast.ai →
Affiliate disclosure: we may earn a commission if you sign up via this link, at no extra cost to you.
💡 Better Stack recommendation: Uptime monitoring, logging, and status pages in one platform. Try Better Stack →
Affiliate disclosure: we may earn a commission if you sign up via this link, at no extra cost to you.