DDevainarihantdeva.hashnode.dev·Jun 6 · 19 min readAgent Harness Comparison: Claude Code, Aider, Cursor Agent, Codex CLI1. Introduction: The Anatomy of Agentic Developer Harnesses Modern software development increasingly relies on language model agents that can read, edit, and execute code without direct human interven10
DDevainarihantdeva.hashnode.dev·Jun 4 · 18 min readContext Window Management: Tactics That Survive Real SessionsThe Illusion of Infinite Context: Effective vs. Nominal Limits Large language models advertise massive context windows, but the practical limit you experience in a real session is often far smaller. T00
DDevainarihantdeva.hashnode.dev·Jun 2 · 18 min readModel Context Protocol Server Design Patterns That Actually Hold Up1. 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 easies30
DDevainarihantdeva.hashnode.dev·Jun 2 · 24 min readHooks vs Skills vs Subagents: Picking the Right Claude Code PrimitiveIntroduction: 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 too20
DDevainarihantdeva.hashnode.dev·Jun 2 · 22 min readToken Economics of Long-Running Agent LoopsThe 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 20
