# Agent Instructions File

> A repository or workspace file (often AGENTS.md) that gives coding agents durable project rules, commands, conventions, and boundaries.

- Category: Context & Memory
- Stage: Rising
- Also called: AGENTS.md, CLAUDE.md, Project memory
- Canonical: https://www.joelcsaji.com/ai-dictionary/agent-instructions-file

## Plain English

It is the project handbook the agent reads before touching the code.

## Example

An `AGENTS.md` tells the agent the project uses Bun, where routes live, and which generated files must not be edited.

## Why it matters

Shared instruction files make multi-tool agent workflows consistent; stale or overloaded instructions become context debt.

## Related

- [Context Engineering](https://www.joelcsaji.com/ai-dictionary/context-engineering)
- [Coding Agent](https://www.joelcsaji.com/ai-dictionary/coding-agent)
- [Agent Skill](https://www.joelcsaji.com/ai-dictionary/agent-skill)

## Mentioned in

- [The Rise of AI Agents: How Autonomous AI Transformed Work in 2025-2026](https://www.joelcsaji.com/blog/rise-of-ai-agents-2025-2026)

## Sources

- [AGENTS.md Convention](https://agents.md/)
- [OpenAI: Codex Agent Loop](https://openai.com/index/unrolling-the-codex-agent-loop/)
- [Claude Code: Extensions](https://code.claude.com/docs/en/features-overview)