The editor war between Emacs and vi users has run since at least 1985, and it was one of the original holy wars on Usenet. Developers now argue about coding agents the way they argued about editors.
An editor changes how a person types code, and an agent writes the code. That moves the skill from keystrokes to instructions.
The CLAUDE.md guide in this topic tests four repositories and finds that Claude Code prefers CLAUDE.md by default and falls back to AGENTS.md. What the agent knows about a project is what a person wrote in those files.
A coding agent is configured through files a person writes, so getting good output from one is a documentation job.
Instruction files
- CLAUDE.md and AGENTS.md: Which File Loads, and a Template. Tests which agent loads CLAUDE.md and which loads AGENTS.md by default, and gives a copyable AGENTS.md template.
Delegation
- Claude Code Subagents: What Each One Sees, With a Tested Example. Shows what a Claude Code subagent starts with and when to pick one over a skill or a hook, with a tested example.
- Agentic Workflows: A Five-Stage Playbook From Prompt to PR. A five-stage playbook that scopes the task before the agent runs and reviews its diff before the change lands.
Memory
- How Memory Works in Claude Code. Explains how Claude Code keeps context across sessions through project files.
- Why Your Coding Agent Keeps Forgetting Everything: Memory Persistence in AI Coding Assistants. Covers the persistence patterns that stop a coding assistant from losing context between sessions.
Choosing a model or tool
- The Best LLMs for Coding in 2026: An Engineering Review. Reviews coding models on Real-SWE 1.1 with the cost per resolved task, and gives a test to run on your own repo.
- Claude Code vs. Gemini CLI for Repository Refactoring. Compares the workflows of Claude Code and Gemini CLI for repository refactoring, with no head-to-head benchmark claimed.
What the code is worth
- Why I rebuilt my best-LLMs-for-coding page. Sets the October jump in coding models beside a benchmark where the best of them passed one staged-change checkpoint in six.
- What I'd still write by hand in a codebase agents can explain. Argues that an agent can explain what code does, while what happened and how it shipped has to be written down by a person.
- Applying Google's pager rule to a build that is always red. Argues that a build light is a pager, and that agent-rate change wears it out unless a team rations what reaches a person.
What does a person have to write for a coding agent?
A person writes the instruction file the agent loads and the prompt file for each subagent. The CLAUDE.md guide tests which file each agent loads, and Claude Code subagents start without the conversation history, so the prompt file has to carry the rules.
Related topics
A coding agent is one agent loop with a harness around it, which the AI agent architecture topic takes apart.
Claude Code's memory is one case of a wider design, and the AI agent memory topic covers the write and read paths behind it.
Instruction files are documentation, and the technical documentation topic covers the page types they sit beside.