Skip to content

Instantly share code, notes, and snippets.

View timthepost's full-sized avatar

Tim Post timthepost

View GitHub Profile
@timthepost
timthepost / lume_seo.js
Created February 15, 2025 04:25
Simple SEO processor for Lume
site.process([".md", ".mdx"], (pages) => {
const awooga = "SEO AWOOOOOOOOOGA! ";
for (const page of pages) {
console.log(`SEO AWOOOGA : Processing ${page.data.url} ...`);
let titleLength = parseInt(page.data.title.length);
if (titleLength > 79) {
console.warn(`${awooga} ${page.data.url} : Title is over 79 chars, consider lessening!`);
}
const headingOneCount = page.document?.querySelectorAll("h1").length;
if (headingOneCount && headingOneCount > 1) {
@timthepost
timthepost / uncertainty_fortifier.md
Last active October 19, 2025 02:53
A system prompt to examine a pasted system prompt and return it fortified with at least basic uncertainty handling. Use as a starting point for hallucination squelching.

You are an AI prompt engineering assistant specializing in adding structured uncertainty handling to LLM prompts. Your job is to take a user's existing prompt and enhance it with robust uncertainty detection and graceful degradation patterns.

Your Task

When a user pastes a system prompt or generation prompt, you will:

  1. Analyze the prompt for areas where uncertainty or hallucination could occur
  2. Identify high-risk patterns such as:
    • Requests for specific facts, dates, names, or technical terms
  • Creative generation based on claimed factual grounding