# How Multi-Step Tasks Are Managed and Why Lists Are Capped at 5 Items in i-have-adhd

> Discover how the i-have-adhd skill manages multi-step tasks with numbered steps and list caps at 5 items. Learn to minimize cognitive load and prioritize effectively for ADHD-friendly task management.

- Repository: [Ayoub Ghriss/i-have-adhd](https://github.com/ayghri/i-have-adhd)
- Tags: internals
- Published: 2026-08-29

---

**The i-have-adhd skill structures multi-step tasks as numbered, bounded steps with explicit progress tracking and caps lists at 5 items to minimize cognitive load, forcing longer sequences into prioritized "now" versus "later" splits.**

The `ayghri/i-have-adhd` repository provides a specialized skill designed to optimize AI interactions for users with ADHD. By enforcing strict structural constraints on multi-step task presentation and limiting list lengths to five items, the system reduces working memory demands and creates clear, actionable roadmaps.

## How Multi-Step Tasks Are Structured

When work requires more than one action, the skill enforces specific formatting rules that prevent cognitive overload.

### Numbered, Bounded Steps

The assistant must produce a **numbered list** where each item represents a single, executable step. This eliminates "and then" chains that force users to hold multiple actions in mind simultaneously. According to [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) lines 42-48, the rule states: "If the work takes more than one step, write a numbered list… Each step is one bounded action."

### State Restatement and Progress Tracking

After completing each step, the assistant explicitly restates the current position and identifies the next action using the format "Step X of Y done: … Next: …". This technique keeps users oriented without requiring them to remember previous context or track their position in a sequence independently.

## Why Lists Are Limited to 5 Items

The hard cap on list length aligns with neuro-cognitive constraints affecting working memory capacity.

### Cognitive Load Management

To prevent overwhelming users, the skill caps any list at **five items**. As defined in [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) lines 103-106, the guideline specifies: "If a list grows past five, split into 'do now' vs 'later', or 'must' vs 'nice to have'… Five items ranked beats ten unranked."

### Prioritization Through Splitting

When tasks exceed five steps, the system mandates splitting items into immediate versus deferred categories. This forced prioritization ensures the most critical actions appear first and prevents presenting users with long, unprioritized sequences that exceed working memory limits.

## Implementation in the Source Code

The constraints are authoritatively defined in the following files:

- **[`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md)** – Contains the numbered step requirements (lines 42-48) and the 5-item cap (lines 103-106)
- **[`README.md`](https://github.com/ayghri/i-have-adhd/blob/main/README.md)** – Explains the overall purpose of the skill and links to the rule set
- **[`AGENTS.md`](https://github.com/ayghri/i-have-adhd/blob/main/AGENTS.md)** – Shows how the skill integrates into various runtimes

## Practical Examples

Below are properly formatted examples demonstrating the 5-item limit and splitting strategy.

### Example 1: Multi-Step Task Within the Limit

```markdown
1. Open `src/auth.ts`
2. Replace the `verifyToken` function (lines 42-58) with the updated snippet
3. Run `npm test -- auth.spec.ts`
4. Verify the login flow works locally
5. Commit the changes with a concise message

```

Each line represents a single, actionable step, and the total count respects the 5-item boundary.

### Example 2: Splitting a Longer Task

```markdown
**Do now (top 5)**
1. Add the new environment variable to `.env.example`
2. Update `src/config.ts` to read the variable
3. Run the unit tests for the config module

**Do later**
4. Update the deployment scripts to set the variable in production
5. Document the change in `README.md`
6. Tag a new release

```

The "Do now" section maintains the cognitive load within acceptable limits, while the "Do later" section preserves important but non-urgent items for future reference.

## Summary

- **Numbered lists** are required for any multi-step task, with each item containing exactly one bounded action
- **State restatement** tracks progress explicitly (step X of Y) to reduce memory demands
- **5-item cap** prevents cognitive overload by limiting working memory requirements
- **Splitting strategy** forces prioritization into "now" versus "later" categories when tasks exceed five steps
- **Source rules** are codified in [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) at lines 42-48 and 103-106

## Frequently Asked Questions

### Why is the limit exactly 5 items?

The limit aligns with research on working memory capacity, often cited as "five plus or minus two" chunks of information. The skill adopts a conservative constraint to ensure users with ADHD are not overwhelmed, following the principle that "five items ranked beats ten unranked" as implemented in the source code.

### How should I handle a task requiring more than 5 steps?

Split the sequence into prioritized categories such as "do now" versus "do later" or "must have" versus "nice to have." This forces ranking of importance and keeps immediate cognitive demands within the 5-item boundary while preserving deferred items for subsequent review.

### What happens if I don't use numbered steps for multi-step tasks?

The skill expects numbered lists for any multi-step work to maintain clarity and reduce working memory load. Without this structure, users must mentally track "and then" sequences, which increases cognitive demands contrary to the tool's accessibility design goals.

### Where are these formatting rules defined in the codebase?

All constraints are authored in [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md), specifically lines 42-48 for step formatting requirements and lines 103-106 for the list length cap. Additional context appears in [`README.md`](https://github.com/ayghri/i-have-adhd/blob/main/README.md), while integration details are documented in [`AGENTS.md`](https://github.com/ayghri/i-have-adhd/blob/main/AGENTS.md).