How the I‑Have‑ADHD Skill Handles Tangents: The Suppress Tangents Rule Explained

The I‑Have‑ADHD skill enforces a strict "Suppress tangents" protocol that isolates secondary topics into optional follow‑up questions, ensuring ADHD users receive single‑threaded, actionable responses without cognitive overload.

The ayghri/i-have-adhd repository provides a conversational skill specifically designed to maintain focus for users with ADHD. Unlike standard AI assistants that may overwhelm users with multiple simultaneous threads, this skill implements a disciplined approach to handling tangents through structured response protocols defined in its core configuration files.

The Suppress Tangents Protocol

The skill's behavior is governed by the "Suppress tangents" rule documented in [skills/i-have-adhd/SKILL.md](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md#L64-L71) (lines 64‑71). This rule prevents the assistant from mixing unrelated topics within a single response, instead enforcing a sequential workflow that respects cognitive load limitations.

Step 1: Isolate the Primary Issue

When generating a reply, the skill never mixes unrelated topics. If a secondary problem is discovered during the generation of a primary answer, it explicitly defers that discovery until the primary issue is fully resolved. This ensures the user sees exactly one clear action per turn.

Step 2: Offer Secondary Issues as Optional Follow‑ups

After completing the main answer, the skill presents any additional discovered topics as optional follow‑ups. The protocol uses a specific phrasing pattern: "Separately: … Would you like me to address that next?" This technique surfaces tangents without forcing them on the user.

Step 3: Handle Mid‑Work Questions as Non‑Tangents

If a question naturally arises while solving the current task, the skill answers it immediately and folds the result back into the ongoing flow. Only if the question truly requires a new thread does it surface it at the end using the optional follow‑up pattern.

Code Examples: Before and After

The repository demonstrates this behavior through concrete examples showing how the same information is structured differently when the suppress tangents rule is applied.

Non‑Compliant Response (Tangent‑Rich)

Here's the fix for the auth bug. By the way, your README is outdated, and you should also upgrade jsonwebtoken.

Compliant Response (Suppress Tangents Applied)

Here's the fix for the auth bug.
Separately: your README is outdated and the jsonwebtoken package is stale. Want me to handle those next?

Handling Integrated Questions

Step 2: Replace `verifyToken`.
Question: "What does verifyToken do?" → It validates the JWT and returns the user ID.
Now continue with step 3...

Configuration Files and Manifests

The tangent suppression behavior is defined across multiple files in the repository:

• skills/i-have-adhd/SKILL.md – Contains the core rule definitions, specifically the "Suppress tangents" section at lines 64‑71
• README.md – Public documentation summarizing the tangents rule in section 4
• GEMINI.md – Runtime‑specific manifest that imports the same rule set for the Gemini model integration

Summary

• The I‑Have‑ADHD skill uses a "Suppress tangents" rule to maintain single‑threaded conversations
• Secondary issues are deferred and offered as optional follow‑up questions using the "Separately:" pattern
• Mid‑work questions that support the current task are answered immediately and integrated into the flow
• The protocol is defined in SKILL.md and referenced in README.md and GEMINI.md
• This approach prevents cognitive overload by ensuring users see one clear action per response

Frequently Asked Questions

Where is the "Suppress tangents" rule defined in the I‑Have‑ADHD repository?

The rule is explicitly defined in [skills/i-have-adhd/SKILL.md](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md#L64-L71) at lines 64‑71. This file serves as the primary skill definition, and the rule is also referenced in the repository's README.md (section 4) and the GEMINI.md runtime manifest.

How does the skill distinguish between a tangent and a relevant mid‑work question?

The skill treats a question as non‑tangent if answering it immediately supports the current task flow. In this case, it provides the answer and continues with the original steps. Only questions that require spawning a new, unrelated thread are treated as tangents and deferred to the end of the response with an optional follow‑up offer.

What specific phrasing does the skill use when deferring tangents?

When deferring secondary issues, the skill uses the pattern: "Separately: [description of issue]. Would you like me to address that next?" or variations like "Want me to handle those next?" This phrasing makes the additional work explicitly optional and clearly separates it from the primary response.

Why does the I‑Have‑ADHD skill suppress tangents instead of handling everything at once?

According to the source code documentation, this design prevents cognitive overload for users with ADHD. By ensuring each turn contains a single, clear action without losing context to digressions, the skill respects the executive function challenges associated with ADHD, making technical assistance more accessible and actionable.

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