i-have-adhd Pre-Send Check Requirements: The 5-Step Validation Rules

The i-have-adhd skill enforces five mandatory pre-send checks—removing opening announcements, deleting closing questions, stripping sidebars, eliminating hedging adverbs, and replacing idioms with literal text—to ensure every response begins with an actionable command and ends with a concrete result.

The ayghri/i-have-adhd repository implements a specialized LLM skill that polishes assistant outputs for clarity and executive function support. According to the source code in [skills/i-have-adhd/SKILL.md](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) (lines 30-41), the pre-send check requirements form a strict validation pipeline applied immediately before final output delivery.

The Five Pre-Send Check Requirements

The checklist defined in SKILL.md consists of five concrete editing operations that must be applied in order:

• Delete the opening sentence if it merely announces what the assistant is about to do (e.g., "Let me show you how" or "Here's the solution").
• Delete the closing sentence if it asks "anything else?" or recaps what just happened, ensuring the final line is a concrete next step rather than a vague wrap-up.
• Remove "by the way" sidebars to eliminate tangential information that can distract readers with limited working memory.
• Strip hedging adverbs such as perhaps, might, or could possibly that add no information and create false uncertainty.
• Replace idioms and figurative phrases (e.g., circle back, get the ball rolling, on the same page) with explicit, literal actions.

These deletions and replacements ensure that every phrase conveys an executable action without cognitive overhead.

Final Verification Criteria

After applying the five requirements, the response must satisfy two final validation questions before transmission:

1. Does the first line tell the reader exactly what to do next?
2. Does the last line clearly state what just happened?

If both answers are affirmative, the response passes the i-have-adhd pre-send validation and is emitted to the user.

Implementing the Pre-Send Checks in Python

Below is a reference implementation demonstrating how to programmatically enforce these requirements using regular expressions:

import re

PRE_SEND_CHECKS = [
    # 1. Remove announcing preamble

    (r'^\s*(Let\'s|Here\'s|I\'m about to) .*?\.\s*', ''),
    # 2. Remove closing question/recap

    (r'\s*(Anything else\?|That\'s it\.)\s*$', ''),
    # 3. Strip "by the way" sidebars

    (r'\bby the way\b[:,]?\s*', ''),
    # 4. Drop empty hedging adverbs

    (r'\b(maybe|perhaps|might|could possibly)\b', ''),
    # 5. Replace idioms with literal wording

    (r'\b(circle back|get the ball rolling|on the same page)\b',
     lambda m: {'circle back': 'return later',
                'get the ball rolling': 'start now',
                'on the same page': 'agree'}[m.group(0)]),
]

def apply_pre_send_checks(text: str) -> str:
    """Run the five pre-send checks on `text`."""
    for pattern, repl in PRE_SEND_CHECKS:
        text = re.sub(pattern, repl, text, flags=re.IGNORECASE)
    # Trim excess whitespace

    return text.strip()

Example usage demonstrates the transformation:

raw_reply = """
Let me show you how.
1. Open src/auth.ts
2. Replace verifyToken
Run npm test.
Anything else?
"""

clean_reply = apply_pre_send_checks(raw_reply)
print(clean_reply)

# → "1. Open src/auth.ts\n2. Replace verifyToken\nRun npm test."

The cleaned output now satisfies the i-have-adhd criteria: the first line is an actionable step, and the last line is the final action itself.

Key Source Files

The pre-send check requirements are distributed across the following repository files:

• skills/i-have-adhd/SKILL.md — Defines the complete skill specification, including the Pre-send check section (lines 30-41) containing the five editing rules and verification criteria.
• README.md — Provides repository overview and installation instructions for the skill.
• plugin.json — Registers the skill with the host agent, exposing the skill name and metadata to the LLM framework.

Summary

• The i-have-adhd pre-send check requirements enforce five specific text transformations: removing preamble sentences, deleting closing questions, stripping sidebars, eliminating hedging adverbs, and replacing idioms with literal language.
• Validation occurs in skills/i-have-adhd/SKILL.md (lines 30-41), with final output verified against two questions regarding the first and last lines of the response.
• Implementation uses ordered regex replacements to ensure the first line is an actionable command and the last line represents a concrete result.

Frequently Asked Questions

What are the five pre-send check requirements in i-have-adhd?

The five requirements are: delete opening sentences that merely announce intent, remove closing sentences that ask "anything else?" or recap actions, strip "by the way" sidebars, eliminate hedging adverbs like perhaps or might, and replace figurative idioms with literal actions. These rules are defined in the Pre-send check section of SKILL.md.

Where are the pre-send check rules defined in the repository?

The pre-send check rules are explicitly defined in [skills/i-have-adhd/SKILL.md](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) at lines 30-41. This file serves as the skill's manifest and contains the complete validation specification.

How does i-have-adhd verify responses after applying the pre-send checks?

After processing the five editing requirements, the skill verifies the response against two final questions: whether the first line tells the reader exactly what to do next, and whether the last line clearly states what just happened. Both must be answered affirmatively for the response to be sent.

Why does the i-have-adhd skill remove hedging adverbs and idioms?

Hedging adverbs are removed because they add no substantive information while creating false uncertainty, whereas idioms are replaced with literal text to ensure every phrase conveys an explicit, executable action that requires no interpretation. This optimization supports readers with limited working memory or executive function challenges.

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