Laravel Code Reviewer — Free

    1

    Technical code reviewer for Laravel and PHP focusing on security, authorization, and Eloquent performance.

    Free

    1 useSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Laravel Code Reviewer — Free

    Example session with this skill installed

    Use Laravel Code Reviewer Free to review this demonstration controller. Review only the supplied code, separate confirmed findings from missing context, and suggest fixes without editing or executing anything.

    <?php
    
    namespace App\Http\Controllers;
    
    use App\Models\User;
    use Illuminate\Http\Request;
    
    class UserDirectoryController
    {
        public function __invoke(Request $request)
        {
            return User::query()->orderByRaw($request->query('sort', 'name'))->paginate(50);
        }
    }
    
    • Read your context and instructions
    • Compiled the laravel code reviewer

    Scope

    Static review of the supplied demonstration UserDirectoryController only. Routes, middleware, the User model, database configuration, and tests were not supplied.

    Confirmed findings

    High — Request input is inserted into a raw SQL ordering expression

    • Location: UserDirectoryController::__invoke(), in the orderByRaw() call.
    • Evidence and explanation: The controller passes the request's sort query parameter directly to orderByRaw(). This allows caller-controlled text to become part of the SQL expression instead of a bound value. The default value, name, applies only when the parameter is absent; it does not validate supplied input. This creates an SQL injection risk. The exact exploit and impact depend on the database and execution environment.
    • Suggested fix: Map a small set of accepted sort options to fixed column and direction pairs, then use orderBy(). For example:
    $sortOptions = [
        'name' => ['name', 'asc'],
        'name_desc' => ['name', 'desc'],
    ];
    
    $sort = $request->query('sort', 'name');
    
    if (!is_string($sort) || !isset($sortOptions[$sort])) {
        abort(422, 'Invalid sort option.');
    }
    
    [$column, $direction] = $sortOptions[$sort];
    
    return User::query()
        ->orderBy($column, $direction)
        ->paginate(50);
    

    Questions and unverified risks

    • Who can access this endpoint? Routes and middleware are needed to assess access controls.
    • Which user attributes are returned? The User model's serialization settings and intended audience are needed to assess possible data exposure.

    Neither issue is confirmed from the controller alone.

    Limits

    This is an author-run demonstration using fixture code, not an independent evaluation. No application code or SQL was executed, and the suggested fix was not runtime-tested. No files were changed. This review does not establish that the surrounding application is secure.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Laravel Code Reviewer — Free guides your AI coding agent through a focused review of user-selected Laravel and PHP files, using the code and context you provide.

    What it reviews

    • Security and validation: untrusted input, mass assignment, output handling, and whether validated data is actually saved.
    • Authorization: visible policies, gates, ownership checks, and supplied route or middleware protections.

    License

    Free for personal use and internal commercial work, including reviewing client code. Private modifications for your own internal use are allowed. Redistribution, sharing, publication, sublicensing, or sale of this package or modified versions requires prior written permission from Anas Shaath. These restrictions do not apply to your reviewed code or review reports. See LICENSE.md in the package for full terms.

    • Query safety: raw SQL, dynamic sorting, and identifier allowlists.
    • Correctness and readability: control-flow errors, misleading names, and unhandled failure paths.
    • Basic performance: concrete query patterns, including relationship access in loops where loading behavior matters.

    How to use it

    Select the files you want reviewed and provide relevant context, such as routes, Form Requests, policies, models, or tests. Ask your agent to review them using this skill.

    The review stays within your selected files and supplied context. Missing information is requested or clearly identified as uncertain.

    What you receive

    A concise report containing:

    • Review scope and missing context.
    • Confirmed findings with severity, location when available, evidence, explanation, and a suggested fix.
    • Separate questions and unverified risks.
    • Limits of what could be checked.

    Evidence-based review

    Findings require a plausible triggering path and impact supported by the supplied code. A familiar API name alone is not treated as proof of a vulnerability.

    Limits

    This is an AI agent skill, not a standalone scanner. It does not provide a complete repository audit or guarantee that all defects will be found. Basic performance checks use static code evidence, not runtime profiling.

    By default, it reviews without editing code. Fixes require an explicit request. It instructs the agent not to read .env files or expose credentials, and not to execute application code, migrations, dependency scripts, or external side effects without authorization.

    Static review is limited to user-selected files and supplied context. Missing routes, middleware, policies, models, or tests may prevent verification. No runtime profiling or complete repository audit is included. The skill instructs the agent not to read .env files or expose credentials, and not to execute application code or migrations without authorization. Findings and suggested fixes require developer verification; the review may miss defects or report false positives.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 11 days ago

    • Free with an Agensi account

    Permissions

    Read Files

    Reads only user-selected files and supplied review context. No external accounts or network access are required. Review does not require file writes or code execution. Code changes require an explicit user request. The skill instructs the agent not to read .env files or expose credentials.

    Listed11 days ago

    What's inside

    Frequently Asked Questions

    Use an AI coding agent that can read SKILL.md and its referenced Markdown guide. Provide access to the selected files or paste the code and relevant context. Setup depends on your agent; this package does not include a GitHub Actions integration.

    It reviews supplied code for security, validation, authorization, query safety, correctness, readability, and relevant basic performance concerns. Findings require supporting code evidence. Missing context is reported separately rather than treated as a confirmed defect.

    Load the skill using your AI agent's supported skill setup, including SKILL.md and references/free-review-guide.md. This package contains Markdown instructions and requires no additional runtime dependencies. Then select the files to review and provide relevant context.

    The report includes scope, confirmed findings with severity, location when available, evidence, explanations, and suggested fixes. Questions, unverified risks, and review limits are listed separately. Code is edited only when you explicitly request fixes.

    The skill provides a reusable Laravel-focused checklist, an evidence standard, and a consistent report format. It instructs the agent to distinguish supported findings from uncertainty. It does not guarantee higher accuracy than other prompts; results depend on the agent and supplied context.

    The review requires only selected source files and supplied context, with no live database connection. The skill instructs the agent not to read .env files or expose credentials, and not to execute application code, migrations, dependency scripts, or external side effects without authorization. These are agent instructions, not a technical sandbox.

    New in Code Quality & Review