The top ten at a glance

14 Best Coding Agent Prompts for CLAUDE.md and AGENTS.md: the top prompts and what you get from each
PromptWhat you get
copilot-instructions.md Generator From Your Repo's Real Build, Test and Style SetupHas GitHub Copilot inspect your repo and write a .github/copilot-instructions.md with the real build, test and lint commands and checkable conventions, plus up to three path-scoped .instructions.md files.
CLAUDE.md or AGENTS.md Audit: Cut, Rewrite as Checkable Rules, Move to SkillsTurns a pasted CLAUDE.md or AGENTS.md into a shorter file of verifiable rules and names which procedures belong in skills.
Codex Writes an AGENTS.md for This Repository From the Real SetupHas Codex inspect your repository and write a short, accurate AGENTS.md with the real build, test and lint commands, conventions and boundaries, checked by actually running them.
Cursor .mdc Rules to Portable AGENTS.md ConverterConverts your pasted Cursor .mdc rules into a single AGENTS.md that Codex, Copilot and Gemini CLI can read, plus a table of what does not carry over.
Cursor Project Rules Generator for Your Repo's StackHas Cursor inspect your repository and write 3-5 scoped .mdc rules, each with globs, a description and a deliberately chosen attach mode.
Lean CLAUDE.md Generator from Repo InspectionClaude Code inspects your repository and writes a CLAUDE.md under 200 lines with verified commands first, checkable conventions, anti-patterns and the reasons behind key decisions.
Claude Code Hook Writer: Auto-Run Lint, Typecheck or Tests After EditsProduces a PostToolUse hook entry for settings.json, the script it calls and a concrete procedure to test that the check runs after file edits.
Frontend Design Direction Brief So Your Coding Agent Avoids the Generic AI LookWrites a DESIGN.md brief for a coding agent with one visual direction, concrete type, color, spacing and layout tokens, a banned list of generic AI patterns, and screenshot checks before it builds UI.
Custom Slash Command Builder for Repeated Claude Code TasksTurns a task you repeat into a custom slash command file with frontmatter, arguments, restricted allowed tools and a fixed output format.
Checkable Guardrail Rules for a Coding Agent: Blocked Commands, Approvals, Secrets and HooksProduces a set of testable guardrail rules for your coding agent, covering blocked commands, approval gates, secret handling and hook checks.

FAQ

About this list

Most of them have the agent inspect your repository first, so the file lists the build, test and lint commands that really work, the conventions the code follows and the places the agent should leave alone. Some run the commands to check them before writing them down.

There are also prompts that cut an overgrown CLAUDE.md down to rules you can check, convert Cursor rules into one AGENTS.md that several agents read, and set up permission rules, a hook that runs your checks after each edit or a custom slash command. Commit the files with the code, so everyone's agents start from the same instructions.

What is the difference between CLAUDE.md and AGENTS.md?

Both are plain Markdown files in the repository that tell a coding agent how the project works. Claude Code reads CLAUDE.md, while Codex, Copilot, Cursor and several others read AGENTS.md. Many teams keep one file and point the other at it, so the instructions only live in one place.

How long should the file be?

Short. The agent reads it at the start of every session, so each line costs attention. Keep commands, conventions and the few rules that prevent real mistakes, and move long procedures into skills or separate documents that the agent opens only when a task needs them. The audit prompt on this list does that trimming for you.

Should the file be committed to the repository?

Yes, for project instructions everyone shares, so the whole team and any agent in continuous integration follow the same rules. Keep personal preferences and machine-specific paths in a local file that your agent supports and that git ignores. Never put secrets or tokens in either file.