Skip to content

Instantly share code, notes, and snippets.

@sekedus
sekedus / gemini-generate-content-text-generation.md
Created August 3, 2026 09:20
Gemini API - Generating content
@sekedus
sekedus / openai-docs-reasoning.md
Created July 21, 2026 22:03
OpenAI Reasoning Models Documentation

OpenAI Reasoning Models Documentation

Source:


Reasoning models like o3 and o4-mini are LLMs trained with reinforcement learning to perform reasoning. Reasoning models think before they answer, producing a long internal chain of thought before responding to the user. Reasoning models excel in complex problem solving, coding, scientific reasoning, and multi-step planning for agentic workflows. They're also the best models for Codex CLI, our lightweight coding agent.

@sekedus
sekedus / get-screenshot-from-youtube.md
Created July 1, 2026 08:02
Get screenshot from YouTube video at a specific time.

In August 2023, Google released an option to copy and save a clean frame from any YouTube or Google Photos video on the web.

Get screenshot from YouTube video at a specific time

  1. Open a YouTube video in Google Chrome (desktop)
  2. Play any video.
  3. Pause the video at the exact timestamp you want.
  4. Double right‑click on the video player.
    > The first right-click open the YouTube context menu, while the second right-click open the Chrome context menu.
@sekedus
sekedus / claude-code-fix-auto-attach-guide.md
Last active August 15, 2026 09:26
Stop Claude Code Extension from Auto-Attaching Your Open Files (VS Code, VSCodium, Cursor).

Claude Code Extension: Fix Auto-Attach

Every time you open a new chat in Claude Code extension, it automatically attaches your currently-open file as context.

This wastes tokens, can leak sensitive files (like .env), and adds noise to every prompt.

This guide shows you how to turn off that feature.


@sekedus
sekedus / README.md
Last active August 9, 2026 09:40
Caveman

Caveman

Turn your AI assistant into a ‘smart caveman’ – terse and efficient – that cuts fluff but keeps all technical accuracy.
Ideal for developers who want concise, no-nonsense answers without losing substance.


Note

If your goal is to reduce token usage, keep in mind that techniques or skills like “Caveman” only shorten an AI’s replies.
They do not reduce the input tokens consumed when the AI processes files or recalls conversation history, which is often the largest cost.

(function(e) {
function a(a) {
for (var t, i, s = a[0], u = a[1], c = a[2], d = 0, l = []; d < s.length; d++) i = s[d], Object.prototype.hasOwnProperty.call(o, i) && o[i] && l.push(o[i][0]), o[i] = 0;
for (t in u) Object.prototype.hasOwnProperty.call(u, t) && (e[t] = u[t]);
m && m(a);
while (l.length) l.shift()();
return r.push.apply(r, c || []), n()
}
function n() {
<!DOCTYPE html>
<!-- saved from url=https://web.archive.org/web/20130713072921/https://arrahmah.com/jadwal-shalat/ -->
<html lang="id">
<head>
<meta charset="utf8">
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Jadwal Shalat</title>
<link rel="stylesheet" href="./style.css">
</head>
<body>
@sekedus
sekedus / getMatchedCSSRules.js
Last active June 11, 2024 12:57 — forked from ydaniv/mozGetMatchedCSSRules.js
Alternative to deprecated window.getMatchedCSSRules
/**
* Alternative to deprecated window.getMatchedCSSRules()
*
* Usage: getMatchedCSSRules(document.getElementsByClassName('my-element')[0]).cssText;
*/
if (typeof window.getMatchedCSSRules !== 'function') {
var ELEMENT_RE = /[\w-]+/g,
ID_RE = /#[\w-]+/g,
CLASS_RE = /\.[\w-]+/g,
ATTR_RE = /\[[^\]]+\]/g,
@sekedus
sekedus / git.md
Last active April 6, 2026 08:27
Git tutorial for beginners