How Rule 3 "End with One Concrete Next Action" Helps Users with ADHD

Rule 3 of the i‑have‑adhd skill requires every response to finish with a single, clearly defined action that takes under two minutes to complete, directly compensating for ADHD‑related working‑memory limitations and task‑initiation barriers.

The ayghri/i-have-adhd repository provides a model‑behavior skill designed to shape AI outputs for readers with ADHD. By enforcing specific response structures in skills/i-have-adhd/SKILL.md, the system reduces cognitive load and helps maintain focus. Rule 3—requiring a concrete next action at the end of every reply—is the critical final step that bridges information delivery to task execution.

Understanding Rule 3 in the i‑have‑adhd Skill

Rule 3 is explicitly defined in skills/i-have-adhd/SKILL.md at lines 57‑63: the model must always end with a single, clearly defined action that the reader can complete in under two minutes. This constraint transforms generic advice into an executable micro‑task, eliminating the ambiguity that typically stalls progress for ADHD users.

Unlike standard conversational AI that may close with open‑ended questions or broad suggestions, this rule mandates a specific behavioral contract. The response architecture follows a consistent "lead‑with‑action → numbered steps → end‑with‑action" pattern (lines 31‑40 and 42‑55), ensuring that every interaction concludes with an immediate, bounded objective.

Core ADHD Challenges Addressed

The rule targets three specific cognitive obstacles documented in the skill’s design rationale:

Working Memory Overload

People with ADHD often experience severe constraints on working memory, capable of holding only a few items simultaneously (referenced at lines 25‑26). By restricting the final output to one tiny step, Rule 3 removes the burden of retaining multi‑step plans or complex context. The reader sees the next move clearly without needing to memorize previous instructions.

Task Initiation Difficulty

Starting tasks requires disproportionate executive effort for ADHD users, making the first action the primary bottleneck (line 27). Presenting a single, ready‑to‑execute command—such as "Run npm test and paste the first failing line"—reduces start‑up friction. The decision is pre‑made, and the action is immediately accessible.

Dopamine Feedback Loops

ADHD brains require frequent, visible rewards to sustain motivation (lines 29‑30). Completing a sub‑two‑minute action delivers an instant sense of accomplishment, triggering dopamine release that reinforces continued engagement. The rule guarantees these micro‑wins occur at the end of every interaction, creating positive momentum for subsequent tasks.

Practical Implementation Examples

The transformation becomes clear when comparing responses before and after applying Rule 3.

Before applying Rule 3:

Here’s how to fix the auth flow. First, open the file, locate the verifyToken function, replace the implementation, then run the test suite. Let me know if you’d like more detail.

After applying Rule 3:

1. Open `src/auth.ts` and locate `verifyToken`.
2. Replace its body with the snippet below.
3. Run `npm test -- auth.spec.ts`.

**Next:** Run `npm test` and paste the first failing line (under 2 min).

The second example demonstrates the rule’s requirements: numbered steps provide the "how," but the final line isolates one concrete action (Run npm test...) with a clear time boundary. This format appears in every response shaped by the skill, as implemented in the repository’s source logic.

Architecture and Consistency

Rule 3 does not operate in isolation. According to the source code in SKILL.md, every response follows a rigid structure: open with an action statement, deliver numbered or bulleted guidance, and close with the mandatory next action. This predictability (lines 31‑55) allows users to anticipate exactly where to find their instruction, reducing the cognitive search cost that often derails focus.

The README.md explains the high‑level purpose of this ADHD‑friendly output style, while INSTALL.md documents how to enable the skill in a runtime environment. Together, these files ensure that Rule 3 remains active across all model interactions, providing consistent support for executive function challenges.

Summary

• Rule 3 in skills/i-have-adhd/SKILL.md (lines 57‑63) mandates that every response end with a single, sub‑two‑minute actionable step.
• This design directly compensates for working‑memory limitations, task‑initiation friction, and dopamine‑feedback requirements specific to ADHD.
• The "end‑with‑action" pattern is part of a broader architectural strategy (lines 31‑55) ensuring predictable, scannable output.
• Practical examples show how vague instructions transform into immediate micro‑tasks that generate quick wins and maintain momentum.

Frequently Asked Questions

Why is a two‑minute time limit specified in Rule 3?

The two‑minute constraint aligns with the ADHD brain’s need for immediate feedback and manageable task chunks. According to the skill documentation, actions that take longer risk triggering procrastination or overwhelm, while sub‑two‑minute tasks provide instant completion gratification that sustains motivation.

How does Rule 3 interact with other rules in the i‑have‑adhd skill?

Rule 3 functions as the terminal component of a three‑part architecture: lead‑with‑action, numbered steps, and end‑with‑action (as defined in SKILL.md lines 31‑55). This sequence ensures that users begin with context, receive structured guidance, and finish with a concrete executable, creating a complete cognitive loop that supports executive function.

Can Rule 3 be applied to technical documentation outside of AI prompts?

Yes. Developers can adopt this principle in technical writing by concluding every section, README, or tutorial with a specific command or file to modify. The repository’s SKILL.md demonstrates that this pattern reduces decision paralysis in any context where ADHD readers might struggle with open‑ended next steps.

Where can I enable the i‑have‑adhd rules in my own AI system?

The INSTALL.md file in the ayghri/i-have-adhd repository provides configuration instructions for activating the skill in compatible runtimes. Once enabled, the system automatically enforces Rule 3 and the associated response patterns across all generated content.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}