agentleFS
vishalguptaxvishalguptax/agent-engineering-kit/site/llms.txt

agent-engineering-kit / site

llms.txt0 starsChanged 24 days agoMITView on GitHub

An open-source (MIT) kit of rules, sub-agents, skills and hooks that gives AI coding agents the habits of a careful senior engineer: plan first, make surgical changes, verify with evidence, simplify and review. A safe installer (npx agent-engineering-kit) puts it into any project in the format of each of 18 AI coding tools (Claude Code, OpenAI Codex, Cursor, GitHub Copilot, Gemini CLI and more), or any agent that reads AGENTS.md. The installer previews every change as a diff before writing,…

  • Reads credentials
Files
llms.txt
# Agent Engineering Kit

> An open-source (MIT) kit of rules, sub-agents, skills and hooks that gives AI coding agents the habits of a careful senior engineer: plan first, make surgical changes, verify with evidence, simplify and review. A safe installer (`npx agent-engineering-kit`) puts it into any project in the format of each of 18 AI coding tools (Claude Code, OpenAI Codex, Cursor, GitHub Copilot, Gemini CLI and more), or any agent that reads AGENTS.md.

The installer previews every change as a diff before writing, merges AGENTS.md, CLAUDE.md and settings files instead of overwriting them, backs up files it changes, records what it did, and can update or uninstall cleanly. The always-on rules in AGENTS.md cost about 600 tokens per session; the full rulebook is reference that skills and agents open one section at a time.

## Pages

- [Home](https://agentengineeringkit.vishalg.in/): What the kit is, the six-step workflow, the parts, supported tools, the installer's safety, and an FAQ.
- [Rules](https://agentengineeringkit.vishalg.in/rules/): The always-on rules block in AGENTS.md, word for word, and the ten sections of the rulebook.
- [Agents](https://agentengineeringkit.vishalg.in/agents/): Each sub-agent: what it does, when it's used, an example, and whether it can edit code.
- [Skills](https://agentengineeringkit.vishalg.in/skills/): Each skill (feature, quick-fix, verify-change, ship, learn, project-conventions): what it does, when to use it, and what it needs.
- [Safety](https://agentengineeringkit.vishalg.in/safety/): The format-on-edit hook, the secret guard and the pre-commit checks, with what each AI tool supports.
- [Tools](https://agentengineeringkit.vishalg.in/tools/): A per-tool table of rules, skills, agents, format on edit and secret guard, with the reason and docs for each.
- [Install](https://agentengineeringkit.vishalg.in/install/): The npx command, the install-with-your-agent prompt, terminal and CI usage, global installs, flags, update and uninstall.
- [How it works](https://agentengineeringkit.vishalg.in/how-it-works/): Where files go, how existing files are merged, backups, the install record, update, uninstall and the safety checks.
- [Guides](https://agentengineeringkit.vishalg.in/guides/): Index of the guides.
- [AGENTS.md vs CLAUDE.md: which file each AI tool reads](https://agentengineeringkit.vishalg.in/guides/agents-md-vs-claude-md/): Which rules file each AI coding tool reads (AGENTS.md, CLAUDE.md, GEMINI.md, copilot-instructions.md, Cursor rules), the tool-specific catches, and how to keep one source of truth.
- [Stop AI coding agents reading your .env](https://agentengineeringkit.vishalg.in/guides/stop-ai-agents-reading-env/): How to keep each AI coding tool out of .env files and secrets: Claude Code deny rules and what they don't cover, ignore files for Cursor, Gemini CLI and others.
- [Format every file your AI agent edits](https://agentengineeringkit.vishalg.in/guides/format-on-edit-hooks/): A format-on-edit hook for Claude Code, Cursor and Windsurf that runs the project's own formatter (Prettier, Biome, ruff, gofmt, rustfmt, …) after every edit.
- [How to stop AI coding agents writing slop](https://agentengineeringkit.vishalg.in/guides/stop-ai-slop/): What AI slop looks like in code, and the habits and checks that prevent it: plan first, surgical changes, verification with evidence, a simplify pass and focused reviews.
- [Stop AI agents weakening your tests](https://agentengineeringkit.vishalg.in/guides/stop-ai-weakening-tests/): How to stop AI coding agents deleting, skipping or loosening tests: an always-on rule, the test-analyzer reviewer, and a pre-commit check that refuses changes to protected test files.
- [Make AI agents prove their code works](https://agentengineeringkit.vishalg.in/guides/verify-ai-generated-code/): A verification loop for AI coding agents: real format, lint, type-check, test and build commands, exercising the change, fixing from the root cause, and reporting what wasn't verified.
- [Review AI-written code with focused agents](https://agentengineeringkit.vishalg.in/guides/ai-code-review-agents/): Code review with focused sub-agents for AI-written changes: test-analyzer, silent-failure-hunter, security-reviewer and frontend-reviewer, when each runs and how tools without sub-agents use them.
- [The best setup for each AI coding tool](https://agentengineeringkit.vishalg.in/setup/): Index of the per-tool setup pages.
- [The Claude Code setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/claude-code/): What the kit sets up for Claude Code: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The OpenAI Codex setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/codex/): What the kit sets up for OpenAI Codex: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Cursor setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/cursor/): What the kit sets up for Cursor: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The GitHub Copilot setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/copilot/): What the kit sets up for GitHub Copilot: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Gemini CLI setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/gemini/): What the kit sets up for Gemini CLI: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Google Antigravity setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/antigravity/): What the kit sets up for Google Antigravity: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Grok Build setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/grok/): What the kit sets up for Grok Build: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Windsurf setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/windsurf/): What the kit sets up for Windsurf: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Kiro setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/kiro/): What the kit sets up for Kiro: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The opencode setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/opencode/): What the kit sets up for opencode: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Kilo Code setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/kilo/): What the kit sets up for Kilo Code: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The JetBrains Junie setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/junie/): What the kit sets up for JetBrains Junie: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Augment Code setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/augment/): What the kit sets up for Augment Code: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Cline setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/cline/): What the kit sets up for Cline: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Zed setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/zed/): What the kit sets up for Zed: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Amp setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/amp/): What the kit sets up for Amp: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Warp setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/warp/): What the kit sets up for Warp: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The Aider setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/aider/): What the kit sets up for Aider: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [The your coding agent setup that ships clean code](https://agentengineeringkit.vishalg.in/setup/generic/): What the kit sets up for your coding agent: rules, skills, subagents, format on edit and secret guard, the files it writes, and how to install it.
- [Changelog](https://agentengineeringkit.vishalg.in/changelog/): Every release, newest first.

## Links

- [GitHub repository](https://github.com/vishalguptax/agent-engineering-kit): source, issues, docs (MIT)
- [npm package](https://www.npmjs.com/package/agent-engineering-kit): `npx agent-engineering-kit`
- [Install with your AI agent](https://raw.githubusercontent.com/vishalguptax/agent-engineering-kit/main/docs/install-with-an-agent.md): the instructions an agent follows to install the kit

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.