Skip to content
Get packs

Guide · Agent skill

How to use the Laravel Verification Loop agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

The laravel-verification skill is a fixed checklist that an AI agent follows on a Laravel PHP project before you open a pull request or deploy. It runs your project's own tools in order, from environment checks through linting, tests, a dependency audit, migration dry-runs and queue status checks, and it stops at the first failing stage. This guide explains what the skill does, how to install it, how to ask for it in plain language, and what to watch out for.

Laravel Verification Loopaffaan-m/ECC · MIT

Quick start

  1. 1Download the ZIP. Download the laravel-verification ZIP from this page. It contains SKILL.md.
  2. 2Unzip into a skills folder. Unzip into ~/.claude/skills/ for personal use, or into .claude/skills/ in your Laravel project for a project-only install.
  3. 3Or upload it in the Claude apps. Go to Settings → Capabilities → Skills in the Claude apps and upload the ZIP.
  4. 4Check project prerequisites. Make sure PHP, Composer, artisan, Pint and PHPStan or Psalm are available in the project.
  5. 5Ask in plain language. Request a check before a pull request or deployment, naming the stage you need.
  6. 6Review failures in order. Fix the first failing stage before continuing, since the skill stops there.

What are agent skills and what does this one do?

An agent skill is a folder containing a SKILL.md file. The file gives an AI agent instructions for a specific task, and the agent reads it when a request matches the task. This skill is for Laravel PHP projects.

When the skill applies, the agent works through these phases in sequence: environment checks (php -v, composer --version, php artisan --version), Composer validation and autoload optimisation, code style with Pint, static analysis with PHPStan or Psalm, tests with coverage, a security audit with composer audit, a migration dry-run with migrate --pretend, cache warm-ups, and queue and scheduler checks. Each layer builds on the one before it, and any failure blocks the release.

  • The skill is written in Spanish and is one of ECC's translated docs.
  • It runs the commands already in your project; it does not bring its own tools.
  • The skill does not need network access.

Install the laravel-verification skill

Download the ZIP from this page. It contains SKILL.md. Install it in one of two ways.

For a personal install, unzip it into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.

  • Personal: unzip into ~/.claude/skills/
  • Project only: unzip into .claude/skills/ inside your Laravel project
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Prepare your Laravel project

The skill assumes your project already has the tools it calls. Check these before you rely on it.

PHP and Composer must be installed. The project needs an artisan file and Pint, plus either PHPStan or Psalm, installed in the project. Coverage uses Xdebug with XDEBUG_MODE=coverage and is optional. If you develop locally with Laravel Sail, the skill can use it. Horizon is only checked if your project uses it.

  • Required: PHP, Composer, a Laravel project with artisan
  • Required in the project: Pint and PHPStan or Psalm
  • Optional: Xdebug for coverage, Laravel Sail, Horizon

How do you trigger the skill with natural requests?

You do not need a special command. Ask the agent in plain language, and mention the event you care about. The skill's description covers environment checks, linting, static analysis, tests with coverage, security scans and deployment readiness, so requests about those topics are the best match.

Good requests name the moment in the workflow, such as before a pull request or before a staging deploy. Be specific about the environment, because some checks behave differently outside production.

  • "Run the laravel verification checks before I open this pull request."
  • "Check whether this Laravel app is ready for a staging deployment."
  • "After this dependency upgrade, run the full Laravel verification loop."
  • "Run the Pint, PHPStan and test steps and tell me where it fails."

Example: a minimal pre-PR run

For a quick check before a pull request, the skill's minimal flow runs the version checks, Composer validation, Pint, PHPStan, tests, the audit, a migration dry-run, the config cache warm-up and the failed-jobs check. Each command is taken from the skill itself.

For a CI-style pipeline, the skill's example runs Composer validation, autoload optimisation, Pint, PHPStan, tests with coverage, the audit, the migration dry-run, the cache commands and schedule:list in order.

composer validate vendor/bin/pint --test vendor/bin/phpstan analyse php artisan test composer audit php artisan migrate --pretend

Safety notes: scripts and the queue healthcheck

The skill runs scripts in your project, so review what it will execute before you let it run on a machine with production access. It does not use the network.

The active queue healthcheck is the riskiest step. It dispatches a no-op job to a dedicated queue and runs one worker to process it. The skill says this should only run in staging or other non-production environments where processing a test job is safe, and that the queue connection must not be sync. An agent can still misapply it if it misidentifies the environment, so confirm the target before approving.

  • Confirm APP_ENV matches the target before the healthcheck step.
  • Make sure the queue connection is not sync.
  • Do not treat the skill as a substitute for human review of destructive migrations; it only asks the agent to look for them.
php artisan tinker --execute="dispatch((new App\\Jobs\\QueueHealthcheck())->onQueue('healthcheck'))" php artisan queue:work --once --queue=healthcheck

Tips for getting reliable results

Fix failures in the order the skill reports them. Later phases such as tests and migrations assume the earlier checks pass, so a failing Pint or PHPStan run is worth resolving before reading test output.

Run the check on a clean working tree when you can, so failures point to your changes rather than leftover files.

  • Keep Pint and PHPStan clean before running full tests and coverage.
  • Check that migration filenames follow the Y_m_d_His_* format and that down() methods allow rollback.
  • Confirm storage/ and bootstrap/cache/ are writable on the target environment.

What people use it for

  • Running a repeatable checklist before opening a pull request on a Laravel app.
  • Checking a Laravel app after a major refactor or dependency upgrade.
  • Preparing a staging or production deployment with a documented sequence of checks.
  • Setting up a CI-style pipeline that runs formatting, static analysis, tests and audit in order.

Frequently asked questions

Is the Laravel verification skill free?

The source material lists the licence as MIT. Check the upstream ECC repository for the exact licence terms before you redistribute it.

What do I need to run the laravel-verification skill?

You need PHP and Composer, a Laravel project with artisan, and Pint plus either PHPStan or Psalm installed in the project. Xdebug is optional for coverage, and Laravel Sail and Horizon are only used if your project uses them.

Where do I install the laravel-verification skill?

Unzip the downloaded ZIP into ~/.claude/skills/ for personal use or into .claude/skills/ in a project. You can also upload it in the Claude apps under Settings → Capabilities → Skills.

Does the skill work on non-Laravel or non-PHP projects?

No. The commands are Laravel and PHP specific, so the skill is not suitable for other stacks.

Is it safe to run the queue healthcheck?

The skill says the active healthcheck should only run in non-production environments where processing a test job is safe. Confirm the environment and that the queue connection is not sync before approving it.

How does this compare with running the commands by hand?

The skill gives an agent a fixed order and a stop-on-failure rule, so the checks run the same way each time. You still need to read the results and review destructive migrations yourself.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Choose this skill if you work on Laravel projects and want a repeatable pre-PR or pre-release checklist that an agent follows in a fixed order. Pick something else if your project is not PHP or Laravel, or if you need an agent to run live queue checks against production, which this skill says should not happen.

Download the skill