Your LLM Is Not DeterministicRun the same report twice and the second result may reorder sections, rename a heading, or round a number differently. Add more instructions and the prompt starts looking like a brittle program writte24GBISep 26 · 2 min readblog.skopow.ski
Frontmatter FTW: How to Make Agent Skills DiscoverableIf your AI agent loads 30 complete Markdown skill files directly into its system prompt, you're burning 40,000 tokens of input budget before the user even types a single word. Worse, drowning the modeAug 1 · 2 min readblog.skopow.ski
Stop Using Full OAuth Handshakes for Internal Microservice AuthOAuth 2.0 was designed to solve a specific problem: delegating access to third-party applications without sharing credentials over public networks. Somehow, over the last decade, teams started blindly52SJul 25 · 3 min readblog.skopow.ski
MCP vs ACP: Clarifying the Protocol Layer in the Agentic EraThe rapid evolution of the AI developer space has shifted our focus from single-turn propmpts to autonomous, multi-agent workflows. Building a system that scans code, writes a PR, runs tests, and deplJun 27 · 6 min readblog.skopow.ski
A Good Agent Skill Is a Contract, Not a PromptWhen people start using AI coding agents seriously, they usually make the same mistake: every time the agent does something wrong, they add another instruction. One more rule, one more exception, one 4May 2 · 7 min readblog.skopow.ski