agentleFS
MrPunyapalMrPunyapal/laravel-auditor/md/agents.md

laravel-auditor / md

AGENTS.md78 starsChanged 3 months agoMITView on GitHub

Wire Laravel Auditor into Codex, Claude Code, Gemini, Cursor, Copilot, and Laravel Boost.

The file

We split this agents.md into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When Boost is installed, Auditor extends it.

    line 18
  2. Very specific:

    If your project already uses Boost, this is the simplest path.

    line 20
  3. Very specific:

    When Boost is absent, auditor:install --agents=claude_code (or your agent) handles the wiring.

    line 24

≈ “Common” is an estimate: at least 1 in 5 of the 41,581 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
agents.md

Laravel Auditor does not run an audit for you. After install, you open your AI agent and ask it to use the laravel-audit skill.

Audit knowledge lives once under resources/auditor. Agent-specific files are thin adapters that point the agent at that shared knowledge. The same skills, guidelines, and rules work across every supported agent.

Integration modes

Laravel Boost

When Boost is installed, Auditor extends it. Boost consumes Auditor's skills and guidelines directly from resources/boost/. The context tools are also registered inside Boost's MCP server automatically through boost.mcp.tools.include.

If your project already uses Boost, this is the simplest path. Run boost:install or boost:update and everything is wired.

Standalone

When Boost is absent, auditor:install --agents=claude_code (or your agent) handles the wiring. It publishes skills, guidelines, and schemas to .ai/, writes agent adapter files, and registers the MCP server for agents that support it.

Both paths produce the same audit knowledge — they just deliver it differently.

Supported agents

The standalone installer supports eight agents:

AgentGuidelinesSkillsMCP
OpenCodeAGENTS.md.agents/skillsopencode.json
Claude CodeCLAUDE.md.claude/skills.mcp.json
Cursor.cursor/rules/laravel-auditor.mdc.cursor/skills.cursor/mcp.json
GitHub Copilot.github/copilot-instructions.md.github/skills.vscode/mcp.json
Gemini CLIGEMINI.md.gemini/skills—
CodexAGENTS.md.agents/skills.codex/config.toml
JunieAGENTS.md.junie/skills.junie/mcp/mcp.json
ZedAGENTS.md.agents/skills.zed/settings.json

Gemini does not support MCP. All other agents receive MCP registration when the installer runs.

Custom agents

The built-in list stays aligned with Laravel Boost. For any other agent, add an entry under custom_agents in config/laravel-auditor.php instead of waiting for a first-class installer target.

'custom_agents' => [
    'my_agent' => [
        'display_name' => 'My Agent',
        'guidelines_path' => 'AGENTS.md',
        'skills_path' => '.my-agent/skills',
        'mcp_config_path' => '.my-agent/mcp.json',
        'mcp_config_key' => 'mcpServers',
        'detect_files' => [],
        'detect_paths' => ['.my-agent'],
    ],
],

Then wire it like any built-in agent:

php artisan auditor:install --agents=my_agent

You can also put the custom key in laravel-auditor.agents for non-interactive installs, or let detection pick it up from detect_files / detect_paths.

mcp_config_path is optional. The installer only auto-registers MCP when that path is a JSON or TOML file. If the agent uses a different config format, omit the path, install skills and guidelines, and register MCP yourself.

What an adapter file contains

An adapter is a short file that tells the agent where to find the audit skill and guidelines. For example, the CLAUDE.md adapter contains:

# Laravel Auditor

This project uses Laravel Auditor for evidence-based Laravel audits.

When asked to audit, review, or assess this application, use the `laravel-audit` skill in `.ai/skills/laravel-audit` and follow `.ai/guidelines/core.md`.

Do not modify application code during an audit. Prefer deterministic project facts from `php artisan auditor:status`, `php artisan auditor:rules`, and the Laravel Auditor MCP tools.

The adapter does not contain the full skill, guidelines, or rules. It points at the shared copies in .ai/.

Asking the agent to audit

Once the agent is wired, give it a clear instruction:

Use the laravel-audit skill to audit this application. Discover the project first, scope the relevant domains, and report only evidenced findings.

That is enough to start a full audit. Ready-to-use prompts for common scenarios — full audit, quick discover pass, domain-focused audits, filtered verification of a single suspicion, re-audits after fixes, and DSA reviews — live in one place:

See Prompt examples.

Read-only boundary

Auditing must not modify application code. Installation may write Auditor-owned resources (skills, guidelines, adapters, MCP config). It will not overwrite user-owned adapter files or an existing laravel-auditor MCP entry without --force.

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.