Cognitive Science Basis for the "Cap Lists at 5 Items" Rule: Working Memory Limits Explained

The "cap lists at 5 items" rule leverages established working memory research showing that humans can actively process only 4±1 chunks of information at once, making five items the optimal limit for maintaining attention and actionability in ADHD support interfaces.

The ayghri/i-have-adhd repository implements Rule 9 to prevent cognitive overload by restricting output lists to a maximum of five items. This constraint directly applies decades of cognitive psychology research on working memory limitations, ensuring that users with attention deficits can effectively process and act on presented information without experiencing decision paralysis.

The Science Behind Working Memory Constraints

Miller's Law and the 7±2 Benchmark

George Miller's 1956 paper The Magical Number Seven, Plus or Minus Two established that the average adult can hold approximately 7±2 chunks of information in short-term memory. While this defines the absolute capacity of working memory, it represents the upper bound for passive retention rather than active processing.

Cowan's Refinement: 4±1 for Active Processing

Subsequent research by Cowan (2001) demonstrated that when information requires active processing—such as evaluating options or making decisions—effective capacity drops to roughly 4±1 items. This narrower window explains why decision-making interfaces must use stricter limits than the classic seven-item benchmark.

Why Five Items Represents the Cognitive Sweet Spot

Presenting more than five items overwhelms the reader's working memory, causing earlier items to be forgotten or ignored through primacy and recency effects. By capping lists at five items, the skill keeps each chunk within the scientifically observed sweet spot of 4-5 chunks, optimizing readability for users with limited attentional bandwidth.

The practical implementation splits longer collections into "do-now" versus "later" groups, ensuring each subset remains actionable without exceeding cognitive limits.

Implementation in the I-Have-ADHD Skill

The enforcement of Rule 9 occurs in the skill's output generation logic. As defined in [SKILL.md](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md), the system automatically partitions any list exceeding five items.

The following pseudocode from the repository demonstrates this logic:

# Example snippet from the skill (pseudocode)

if len(generated_list) > 5:
  do_now = generated_list[:5]           # first five items

  later = generated_list[5:]            # remaining items

  output = format(do_now, later)        # separate sections in the response

else:
  output = format(generated_list)

When the skill produces a list of actions, it automatically enforces the cap, presenting at most five items before prompting the user to continue with the next set.

The rule configuration appears in both the OpenAI and Gemini agent configurations:

• OpenAI configuration: [agents/openai.yaml](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/openai.yaml)
• Gemini configuration: [agents/gemini.toml](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/gemini.toml)

Summary

• The "cap lists at 5 items" rule derives from empirical research on working memory constraints, specifically Miller's 7±2 benchmark and Cowan's refined 4±1 limit for active processing.
• Five items represents the practical maximum for maintaining attention and ensuring actionability in decision-making contexts.
• The ayghri/i-have-adhd repository implements this constraint in [SKILL.md](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) by automatically splitting lists exceeding five items into sequential "do-now" and "later" segments.
• This approach prevents cognitive overload and supports users with ADHD by respecting the biological limitations of working memory capacity.

Frequently Asked Questions

Why is the limit five items instead of seven?

While Miller's original research suggested 7±2 items for passive short-term memory, Cowan's subsequent findings indicate that active processing—such as evaluating tasks or making decisions—reduces effective capacity to 4±1 items. Five items sits safely within this narrower window, accounting for the additional cognitive load required to act on information rather than merely recall it.

How does the i-have-adhd skill handle lists longer than five items?

According to the source code in [SKILL.md](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md), the skill automatically partitions overflowing lists into separate "do-now" and "later" sections. The first five items appear immediately, while remaining items are deferred to subsequent interactions, ensuring no single view exceeds working memory limits.

Does this rule apply to all types of content or just task lists?

Rule 9 specifically targets actionable lists and decision-making contexts where users must evaluate and prioritize items. The cognitive science basis applies broadly to any interface requiring active processing, though the i-have-adhd implementation focuses on task-oriented outputs to maximize executive function support for users with attention deficits.

What happens if a user has a higher-than-average working memory capacity?

The 4±1 and 7±2 figures represent population averages. While individual capacity varies, designing for the constrained case ensures accessibility across the full spectrum of users, including those with ADHD who experience significantly reduced working memory resources. The i-have-adhd skill prioritizes universal usability over optimizing for high-capacity outliers.

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