Audit report
$250
3 days · 1 revision
You want to know where you stand before you decide what to fix.
- A written report in plain English
- Security, data and deploy risks ranked by how serious they are
- A fix plan, in the order I would do the work
Fixed-price project
I read the code of an app built with an AI tool, find what is unsafe to launch, and rank it by how serious it is. A written report starts at $250 and takes 3 days. Fixes cost more.
Who it is for
Signs
Scope
I search the code and its history for keys, tokens and passwords. I check that secrets live in environment settings and not in files in the repository, and that debug mode is off in production. If a key was ever committed, treat it as exposed and replace it, even if the file was cleaned up later.
A login page does not mean each user sees only their own data. I follow every route and API call that takes an id and check whether the code asks who may see that record. OWASP lists broken access control first in its 2025 Top 10, so I look at this first.
I look at who can read and write each table. If your app talks to a hosted database straight from the browser, I check that row-level rules exist on every table. I also look for missing unique indexes and foreign keys, how data is deleted, and whether development and production use separate databases and keys.
I read how the code handles failures: swallowed errors, no alerts, no health check. OWASP also lists mishandling of exceptional conditions and logging and alerting failures in its 2025 Top 10. If you take payments or use webhooks, I check that the code verifies who sent each request, ignores repeat events and moves slow work to a queue.
I check how the app gets from your laptop to a live server: environment settings, separate keys for production, a way to roll back, a health check, and whether a backup has ever been restored. A backup that nobody has tested is only a hope.
I check whether another developer could work in the code: folder layout, repeated code, outdated packages with known problems, and whether the parts that handle money, access and deletion have any tests. In the larger packages I clean this up and write a handover guide.
Packages
Fixed price in US dollars. Each package lists what is included.
$250
3 days · 1 revision
You want to know where you stand before you decide what to fix.
$900
10 days · 2 revisions
You want the worst problems gone before you launch.
$2,200
21 days · 3 revisions
You want the app cleaned up and ready for real users and for the next developer.
Process
Step 1
You send a GitHub or GitLab invite, or a zip file. You tell me what the app does, which AI tool built it and where you plan to launch. Tell me what worries you most. Keep live passwords and production keys out of what you send.
Step 2
I read the code and check the six areas above, starting with the parts that handle users, money and data. I work from the code, not from your live servers.
Step 3
In Audit and fixes and Production-ready, I fix the most serious problems first. In Production-ready I also clean up the structure and set up a safe deploy, backups and monitoring.
Step 4
You receive a written report that ranks each problem by severity. With the fix packages you also get the changes in your repository. Production-ready adds a handover guide for you and your team.
Proof
Article
Practical AI agent security for Laravel apps and MCP servers. Learn how to limit prompt injection damage with scoped tokens, approvals, and...
Article
How to add AI code review to a Laravel repo, what it catches and misses, and how to pick a tool. Numbers come from papers and project pages.
Article
What tends to break when a vibe-coded app goes live, from access control to secrets and payments, with sources and a short pre-launch checkl...
FAQ
I check six areas: secrets and keys, login and permissions, database rules, error handling, deploy and backups, and code structure. I read the code that handles users, money and data with the most care. I work from the code only, so I do not need access to your live servers.
It is a written document in plain English. Each finding says what is wrong, where it is in the code, why it matters and how to fix it. Findings are ranked by how serious they are, and a fix plan puts them in the order I would do the work.
I need a link to your code (a GitHub or GitLab invite, or a zip file) and a short note on what the app does, which AI tool built it and where you plan to launch it. Tell me what worries you most. Keep live passwords and production keys out of what you send.
A full rewrite, new features and a live attack test of your running site are not included. A review also cannot promise that your app has no flaws. The Audit report package finds and ranks problems but does not fix them. Fixes start with the Audit and fixes package.
No, I will not rewrite your whole app. I start with the risky parts and leave working code alone. If a part cannot be saved, I explain why and quote it before I touch it.
It does not fit if you want new features built, a full redesign or a rewrite in another language. It also does not fit if you cannot share the code. A weekend toy with no real users and no private data may not need a paid audit. Tell me your plan and I will say which package fits.
I review apps from Lovable, Bolt, Cursor, Claude Code, v0, Replit and similar tools. Laravel and PHP are my strongest stacks, and I also review JavaScript and TypeScript apps. Tell me which stack your app uses.
Pick Audit report if you only want to know where you stand. Pick Audit and fixes if you want the most serious problems fixed: keys, login, database rules and errors. Pick Production-ready if you also want a cleaner structure, a safe deploy, backups, monitoring and a handover guide.
If your work does not fit these packages, tell me what you need and I will reply with a plan.
More projects
From $500
An MCP server for your Laravel app, so an AI agent can use only the tools you allow.
From $250
Set up Claude Code as a small team, with your rules, sub-agents and skills.
From $300
Contacts, companies and deals kept in step between Laravel and HubSpot.