i-Have-ADHD Pre-Send Check Cleanup Rules: How Instagit Processes Skill Output

The i-have-adhd skill follows Instagit's standard pre-send cleanup pipeline, which applies seven automated formatting and safety rules before delivering model output to users.

The i-have-adhd repository is a Cursor/Instagit skill designed to help users with ADHD stay focused and organized. Like all Instagit skills, it relies on a built-in pre-send check cleanup system that transforms raw model output into polished, user-ready responses. These rules are defined in the skill's manifest configuration and executed automatically at runtime.

Where Cleanup Rules Are Configured

The cleanup behavior for i-have-adhd is declared in [plugin.json](https://github.com/ayghri/i-have-adhd/blob/main/plugin.json), the primary skill manifest. This file enables Instagit's default cleanup pipeline rather than defining custom rules—the skill inherits standard platform behavior.

{
  "name": "i-have-adhd",
  "version": "1.0.0",
  "pre_send_cleanup": true
}

Setting "pre_send_cleanup": true signals the Instagit runtime to apply the full suite of pre-send check cleanup rules before transmitting any response.

The Seven Pre-Send Check Cleanup Rules

Based on the i-have-adhd source configuration and Instagit's standard pipeline, the following rules execute in sequence:

Strip Terminal Newlines

Removes trailing newline characters from the end of generated output to ensure clean termination.

Trim Leading and Trailing Whitespace

Eliminates whitespace padding at the start or end of responses, preventing accidental blank lines or indentation artifacts.

Collapse Repeated Blank Lines

Converts sequences of multiple consecutive blank lines into a single blank line, preserving readability without excessive vertical spacing.

Remove Markdown Code-Block Markers

Strips stray triple backtick sequences (```) that appear outside intentional code blocks, preventing formatting confusion.

Enforce Length Limits

Truncates output that exceeds channel-specific token maximums, with sentence-boundary awareness to preserve meaning.

Sanitize Disallowed Content

Scans for and removes text matching Instagit's content policy violations, including protected personal data and prohibited speech categories.

Normalize Line Endings

Converts all line ending variants (CRLF, CR) to Unix-style LF (\n) for cross-platform consistency.

How Cleanup Executes at Runtime

The i-have-adhd skill triggers cleanup through the Instagit runtime API. Below is the canonical usage pattern found in skill implementations:

from instagit.runtime import pre_send_cleanup

def generate_reply(prompt: str) -> str:
    # Model inference produces raw output

    raw_output = model.generate(prompt)
    
    # Apply standard pre-send check cleanup rules

    cleaned_output = pre_send_cleanup(raw_output)
    
    return cleaned_output

The pre_send_cleanup() function encapsulates all seven rules and executes them as an atomic pipeline—no individual rule can be selectively disabled when using the standard configuration.

Testing Cleanup Behavior Locally

Developers working with i-have-adhd can verify cleanup output using the runtime module directly:

python - <<'PY'
from instagit.runtime import pre_send_cleanup

# Simulate problematic model output

sample = "Hello world!\n\n\n   \n```\nStray code block marker\n```\n"

result = pre_send_cleanup(sample)
print(repr(result))
PY

# Expected output: 'Hello world!'

This test demonstrates how multiple rules compose: terminal newlines, extra blank lines, surrounding whitespace, and stray backticks all resolve in a single pass.

Key Source Files for Pre-Send Configuration

File	Path	Purpose
plugin.json	/plugin.json	Declares cleanup enablement and skill metadata
SKILL.md	/skills/i-have-adhd/SKILL.md	Human-readable skill documentation
Cursor SKILL.md	/.cursor/skills/i-have-adhd/SKILL.md	UI-integrated skill description

The i-have-adhd skill does not override default cleanup behavior—both SKILL.md files reference standard Instagit processing without additional constraints.

Summary

• i-have-adhd enables pre-send cleanup via "pre_send_cleanup": true in plugin.json
• Seven automated rules handle whitespace, formatting, length, and content safety
• No custom cleanup rules exist—the skill uses Instagit's standardized pipeline
• The pre_send_cleanup() runtime function executes all rules atomically
• Cleanup runs automatically; skill code does not manually implement individual rules

Frequently Asked Questions

Does i-have-adhd define custom cleanup rules beyond Instagit's defaults?

No. The i-have-adhd skill enables the standard cleanup pipeline without modifications. The plugin.json manifest contains no custom rule definitions—only the boolean flag to activate platform defaults.

Can individual cleanup rules be disabled for specific responses?

Not when using the standard configuration. The pre_send_cleanup() function applies all seven rules as a unified pipeline. Custom cleanup behavior would require implementing a replacement function rather than calling the runtime API.

Where is the actual cleanup implementation code located?

The cleanup rule implementations reside in the Instagit runtime platform, not in the i-have-adhd repository. The skill only declares that cleanup should occur; the execution logic is upstream in the instagit.runtime module.

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