Skip to content

Instantly share code, notes, and snippets.

@szymdzum
szymdzum / ARTICLE_MCP_VS_CLI_FOR_AGENTS.md
Created December 1, 2025 10:21
MCP vs CLI: Which Interface Do AI Agents Actually Prefer? - A Benchmark-Driven Comparison of Browser Automation Paradigms

MCP vs CLI: Which Interface Do AI Agents Actually Prefer?

A Benchmark-Driven Comparison of Browser Automation Paradigms


Introduction

When building tools for AI agents, developers face a fundamental interface choice: expose functionality through the Model Context Protocol (MCP), or provide a traditional command-line interface (CLI) that agents invoke via shell commands.

@szymdzum
szymdzum / SELF_DOCUMENTING_SYSTEMS.md
Created December 1, 2025 10:21
Self-Documenting Systems for Autonomous Agents - How tools can teach agents to use them without external documentation
@szymdzum
szymdzum / [slug].txt.ts
Last active November 30, 2025 21:39
llms.txt generator for Astro blogs - zero dependencies
import type { GetStaticPaths } from "astro";
import { siteConfig } from "@/site-config";
import { llmsPost } from "@utils/llms";
import { formatUrl } from "@utils/path";
import { type BlogPost, getAllPosts } from "@utils/posts";
export const getStaticPaths: GetStaticPaths = async () => {
const posts = await getAllPosts();
return posts.map((post) => ({
params: { slug: post.slug },
@szymdzum
szymdzum / glab.md
Last active November 26, 2025 22:00
glab skill
name description
Pipeline Investigation
Debug GitLab CI/CD pipeline failures using glab CLI. Investigate failed jobs, analyze error logs, trace child pipelines, and compare Node version differences. Use for pipeline failures, job errors, build issues, or when the user mentions GitLab pipelines, CI/CD problems, specific pipeline IDs, failed builds, or job logs.

Investigating GitLab Pipelines

Use this skill when investigating GitLab CI/CD pipeline issues.

When to Use

@szymdzum
szymdzum / iam.ts
Last active August 24, 2025 17:39
boltzmann-opus
#!/usr/bin/env node
/**
* Developer Hub MCP Server
*
* THE SCIENCE: Why This Works
*
* 1. SEMANTIC MEMORY ARCHITECTURE
* Human brains have two memory systems:
* - Episodic (specific events)
@szymdzum
szymdzum / claude-memory.md
Last active September 20, 2025 06:42
claude

Claude Code Memory Management & Context Optimization Guide

How Claude Code Memory Works

Claude Code uses a simple but powerful file-based memory system centered around CLAUDE.md files. These are automatically loaded into context when Claude Code starts, providing persistent memory across sessions.

Memory File Hierarchy

~/Developer/

HTML is an Interface, and We've Been Treating It Like a Whiteboard

The Interface We All Agreed to Ignore

Picture this: You're reviewing a pull request and you see this beauty:

interface Product {
  sc_kDDrLX: string;      // ??? maybe a title?
 eYUOdi: any; // your guess is as good as mine