agents.md
A markdown instruction file that agents read to learn project conventions, analogous to CLAUDE.md for Claude.
As defined in Lex Fridman Podcast #501 - DHH
Other definitions
A community standard for agent specifications that reached 60,000 repositories, though limited to packaging components and conventions rather than full harness logic
Rethinking AI Agents: The Rise of Harness EngineeringA markdown file used in some agent frameworks to define system-level instructions and guidance that are automatically provided to the model without requiring the user to repeat them.
Harness Engineering Masterclass: Technical Deep Dive on how to build Agentic Systems
Where it comes up · 3
- Lex Fridman Podcast #501 - DHH2 claims
- Harness Engineering Masterclass: Technical Deep Dive on how to build Agentic Systems1 claim
- Rethinking AI Agents: The Rise of Harness Engineering1 claim
Most important claims · 4 of 4Search all claims for “agents.md” →
- Instructions are the first harness primitive; they tell the model who it is, what work it does, the tone, constraints, coding style, code review rules, and actions to avoid. Examples include agents.md, claude.md, system prompts, custom instructions, repository rules, and cursor rules. Instructions move repeated guidance into the environment so users don't have to repeat them.Harness Engineering Masterclass: Technical Deep Dive on how to build Agentic Systems
- Claude Code's system prompt for Opus 5 shrank by 80% because the agent needed far less human instruction and was being degraded by overly prescriptive human guidance, much like a programmer damaged by a micromanaging boss.Lex Fridman Podcast #501 - DHH
- When Anthropic stopped allowing Claude subscriptions to be used in other harnesses like OpenCode, DHH saw it as a protectionist move to lock users in, and noted Claude refuses to read AGENTS.md and other standard instruction files.Lex Fridman Podcast #501 - DHH
- Standards emerged in agent harness design: Agents.md reached 60,000 repositories and Anthropic's agent skills added reusable procedures, but both packaged components, conventions, and snippets rather than the full harness itself.Rethinking AI Agents: The Rise of Harness Engineering