Free AGENTS.md Generator
An AGENTS.md generator turns your repo stack, commands, conventions, testing rules, and review preferences into a ready-to-paste instruction file for Codex and other coding agents. Put it at the root of your repo so every agent session starts with the same project-specific guidance.
Repository
Commands
Agent instructions
AGENTS.md
62 lines · updates as you type
# AGENTS.md ## Acme Platform Multi-tenant customer dashboard for onboarding, billing, and workflow automation. ## Stack - Stack: Next.js App Router, TypeScript, Prisma, Postgres, Tailwind CSS - Package manager: pnpm ## Commands - Development: `pnpm dev` - Build: `pnpm build` - Tests: `pnpm test` - Lint: `pnpm lint` - Typecheck: `pnpm tsc --noEmit` ## Repository Layout src/app: Next.js routes and server actions src/components: shared React components src/lib: business logic and service clients prisma: schema, migrations, and seed scripts ## Coding Conventions - Use TypeScript for all new application code - Prefer server components unless interactivity requires "use client" - Keep edits scoped to the requested behavior - Match existing naming, import aliases, and styling patterns ## Testing Expectations - Run lint and typecheck before final handoff - Add focused tests when changing business logic - For UI changes, verify the affected route in a browser-sized viewport ## Agent Workflow - Read nearby files before editing - Reuse existing helpers before adding new abstractions - Explain any skipped validation clearly in the final response - Preserve unrelated user changes in the worktree ## Review and Handoff - Call out risky migrations, auth changes, and billing changes - Include exact files changed and validation commands run - Prefer small, reviewable patches over broad refactors ## Safety Rules - Do not run destructive git commands unless explicitly requested - Do not commit secrets, tokens, .env files, or local credentials - Do not change generated files by hand unless the generator is unavailable --- _Generated with the free [AGENTS.md Generator](https://www.tornic.dev/tools/agents-md-generator) by Tornic._
How to create an AGENTS.md file
- Describe the repository. Enter the project name, purpose, stack, package manager, and the key folders an agent should understand before editing.
- Add exact commands. Paste the real dev, build, lint, typecheck, and test commands. Exact commands help agents verify changes without guessing.
- Write agent rules. Capture conventions, testing expectations, handoff rules, and safety constraints such as preserving unrelated changes.
- Copy or download AGENTS.md. Copy the generated Markdown or download AGENTS.md, then place it at the repository root before your next agent session.
Frequently Asked Questions
What is AGENTS.md?
+
AGENTS.md is a Markdown instruction file for coding agents. It gives tools like OpenAI Codex repo-specific context: what the project does, how to build and test it, which conventions to follow, and which commands or files require caution.
Where should I put AGENTS.md?
+
Put AGENTS.md at the repository root so coding agents can find it before making changes. For monorepos, you can also add package-level AGENTS.md files when a subproject needs different commands or rules.
What should AGENTS.md include?
+
A useful AGENTS.md includes the project purpose, stack, exact dev/build/test commands, repository layout, coding conventions, test expectations, review rules, and safety constraints such as never editing generated files or committing secrets.
Is AGENTS.md only for OpenAI Codex?
+
No. AGENTS.md is a portable convention for coding-agent instructions. It is especially useful for Codex-style workflows, but the same file can help any assistant or teammate understand how to work inside the repo.
How is AGENTS.md different from CLAUDE.md?
+
CLAUDE.md is Claude Code's native project memory file. AGENTS.md is a more general coding-agent instructions file. Many teams keep both so Claude Code and Codex-style agents receive equivalent project guidance.
Need repeatable agent runs, not just better instructions?
AGENTS.md gives coding agents the rules. Tornic turns those rules into repeatable Claude, Codex, and Cursor workflows so the same checks run the same way every time.
Try Tornic