@arihantdeva

Model Context Protocol Server Design Patterns That Actually Hold Up

21h ago · 18 min read · 1. The Naive Wrapper Anti-Pattern: Why Directly Wrapping APIs Leaks Context Tokens When a service simply forwards a request to an external API and returns the raw response, it appears to be the easies

Hooks vs Skills vs Subagents: Picking the Right Claude Code Primitive

21h ago · 24 min read · Introduction: Extending Claude Code Claude Code gives developers a programmable interface to Claude’s language model. It lets you embed model calls inside ordinary code, combine them with external too

Token Economics of Long-Running Agent Loops

21h ago · 22 min read · The Hidden Cost of Multi-Turn Context Windows When an autonomous agent runs a loop that spans many turns, the model must keep the entire conversation history in its context window. Each new turn adds

How Claude Code's Skills System Actually Works

21h ago · 25 min read · Engineering teams building agentic workflows often struggle with the "last mile" of tool integration. While generic API calls are well understood, the bridge between a language model and a local files

LLM agents in quantitative finance, where they actually pay off

21h ago · 7 min read · The seductive pitch for large language models in finance is that they read 10-Ks fast. The more interesting reality is what they let you do on top of that: build small, disciplined agents that compose