Best Claude Code Skills: Which Ones Are Actually Worth Installing
I tested public Claude Code skills and deleted most of them. Which five are worth installing, which to skip, and why security matters more.
Kyle Redelinghuys
Practical insights on Go, AI, and fintech from someone who builds things that work
Latest posts
AI without a system of record is just chat
Chat assistants forget everything. Building an AI system of record with Notion, Claude Code and written skills, plus the failure modes that shaped
What Actually Goes Wrong When You Give AI Agents Production Access
Replit deleted a production database during a code freeze. Amazon Q shipped a wiper prompt to 964,000 installs. The controls that fix this exist.
The Claude Code Workflows That Actually Survived Contact With Production
Six Claude Code workflows I still use daily after eighteen months, four I abandoned, and the pattern that separates them.
How I Got the UK Global Talent Visa as a Software Engineer (2026 Update)
A 2026 guide to the UK Global Talent visa for software engineers, covering what's changed since the Tier 1, the current criteria, and what evidence actually works.
I Got Tired of Paying for Server Monitoring I Couldn't Afford, So I Built My Own
I built SrvMon because server monitoring is too expensive or too complex to self-host. A 6MB agent, a Go API, and a Vue dashboard - open source
Beyond the Terminal: Using Claude Cowork to Close the Productivity Gap
I'm 50x more productive coding with Claude Code. Here's how I'm using Claude Cowork and a custom MCP server to start closing that gap
Teaching a Transformer to Read DNA: How EabhaSeq Finally Works
Months of failed tests, a breakthrough in reference-conditioned generation, and what happened when synthetic training met real T21 samples.
Claude Code agents and subagents: what they actually unlock
I set up Claude Code agents early, stopped using them, then dug back in. Here's what the subagent architecture actually enables.
Teaching AI Agents to Find Their Own Way Around Cont3xt
AI agents were signing up to Cont3xt autonomously. Here's how I built self-registration, skill.md discoverability, and a governance layer.
The Claude Agent SDK: What It Is and Why It's Worth Understanding
The Claude Agent SDK exposes the infrastructure behind Claude Code as a lib. Here's what subagent orchestration and session persistence look like
I Built a Cost Tracker for Claude Code to See If My Subscription Was Worth It
I built a Go binary to track my Claude Code API spend. The answer: $1,428 in API equivalent costs on a $200/month Max subscription.
Claude Code Pricing Guide: Which Plan Actually Saves You Money
I tracked 10 billion tokens across 8 months of Claude Code usage. The API would have cost $15,000. I paid $800 on Max.
OpenClaw: How I Built a Personal AI Operations Centre with Telegram and a $6 Droplet
Running OpenClaw on a $6 droplet via Telegram. It handles my blog analytics, email outreach, research and even built itself voice transcription.
Claude Code Hooks: A Complete Guide to Automating Your AI Coding Workflow
Learn how Claude Code hooks give you deterministic control over your AI agent. Covers lifecycle events, three handler types, practical examples.