David Montgomery DMontgomery40

DOCX ↔ PDF Round-Trip Normalization Prompt

For use with GitHub Copilot, Claude, or other AI coding assistants

Objective: Eliminate DOCX ↔ PDF round-trip layout drift through deterministic normalization.

Normalization Pipeline:

Ralph Audit Agent Instructions (OpenAI Codex)

Safety Notice (Customize)

If this codebase is production, handles money, or touches sensitive data: treat this audit loop as a high-risk operation. Run with least privilege, avoid exporting long-lived credentials in your shell, and keep the agent in read-only mode.

Step 0 before trying to fix your vibe slop

Important: this is not about writing prettier logging copy. The point is to make operatorHint act like a tiny implementation diagnosis that has already done some of the debugging work for the next AI agent. The strongest mode is not supposed to produce vague junk like look at notifications. It is supposed to produce a compact root-cause note, more like PATCH /notifications -> parseNotificationPatchBody(); missing body.text guard before schema parse or RunPage reconnect effect opens a new EventSource before cleanup. In other words: route, function, boundary, missing guard, broken invariant, likely culprit. Not decorative logging.

Weak vs Strong

These are real runs against the same tiny repo, same file, same error path.

Weak:

Ragweld Codex Automation Hardening Report

Last updated: March 10, 2026

Incident summary

This report documents why the existing Codex Desktop automation path for ragweld was treated as unsafe, what concrete failures we observed, what we changed locally to stop the lying/fake-green behavior, what we changed in the repo to harden the worker lanes, and what still remains open.

Primary trigger session:

name	pr-loop
description	Run the full GitHub pull request loop for the current branch, including review waits, fix and amend cycles, merge, and post-merge verification.

PR Loop

Overview

Run the whole PR lifecycle without dropping the follow-through steps. Use this skill when the user wants a repeatable "open PR, wait, inspect review, fix, recheck, merge, verify deploy" loop for the current branch.

name	thorough-code-review
description	Use when the user wants a serious, bug-focused code review of a repo, branch, PR, or diff and explicitly wants more than the default handful of findings. Focus on correctness, regressions, missing tests, fake or gamed tests, security, portability, cleanup, and operational risk.

Thorough Code Review

When To Use

Use this skill when the user asks for any of these:

name	frontend-code-review
description	Use when the user wants a frontend-focused, bug-first review of UI code or tests, especially in TypeScript or JavaScript React and Vite apps, with light Streamlit coverage. Focus on concrete correctness issues in rendering, state, interactions, async flows, accessibility, visual behavior, and fake or weak frontend tests.

Frontend Code Review

When To Use

Use this skill when the user asks for any of these:

	class ExtensionToggler extends ScryptedDeviceBase implements Settings, OnOff {
	async getSettings(): Promise<Setting[]> {
	return [
	{
	key: 'devices',
	type: 'device',
	title: 'Devices',
	description: 'The devices on which the extension will be toggled.',
	multiple: true,
	value: this.getJSON('devices'),

	const os = require('os');
	const { execSync } = require('child_process');

	const mo = await mediaManager.createMediaObjectFromUrl('https://user-images.githubusercontent.com/73924/230690188-7a25983a-0630-44e9-9e2d-b4ac150f1524.jpg');
	const image = await mediaManager.convertMediaObject<Image & MediaObject>(mo, 'x-scrypted/x-scrypted-image');
	const detectors = [
	'@scrypted/openvino',
	// Uncomment other detectors if available
	// '@scrypted/coreml',
	// '@scrypted/onnx',