[email protected]

Have your Claude Code app audited

You built your application with Claude Code and want to know whether it can go live. We review code, data access and operations and attack a test instance. Confirmed findings come with proof. Every finding carries one of four states: confirmed, refuted, suspicion without confirmation, or not tested.

Reviews
Cloud Wolke HerausforderungServer NachhaltigkeitIT Beratung Service Consulting SoforthilfeTimo Wevelsiep Robin ZinsExperten für Innovation Migration AWSHetzner Hosting zuverlässig

Customers from mid-sized business, industry and the public sector, in Germany and internationally

  • Stadtwerke Brühl
  • DGHO e.V.
  • solarwerkstatt berlin
  • Krüger Dirndl
  • ABCO Water Systems
  • Golem.de
  • EVADXB
  • nextGYM
  • AInergy
  • ml&s
  • Odiseo Solutions
  • Annota
  • ARGE
  • SweetConnect GmbH
  • Aphy AG
  • CORGOS
  • Rekorder
  • SolidProof
  • Yonju
  • Keymate
  • Paritel
  • Mr. Clipart
  • Millenium
  • Negosh
  • Führerscheinmacher
  • BTI
  • Deutsche Gesellschaft für Datenschutz
Read client reviews

Audit focus

What we look at closely in Claude Code apps

Claude Code is an assistant from Anthropic that writes code in a repository from the terminal, an IDE or the cloud, runs commands and creates commits. What it produces lives in the operator's repository, together with the settings and instruction files of the assistant. For a security audit, what lands in the code and which permissions the assistant had while doing it are what matter.

  • Access checks per record and per tenant in the generated routes
  • Commands, file access and outgoing requests in the generated code
  • Secrets and keys in the repository and its history
  • Dependencies, build and deployment through to production

Claude Code in detail

Auditing an app built with Claude Code: what we look at

According to the Anthropic documentation, the user is responsible for reviewing proposed code and commands before approving them (as of November 2026). We therefore audit the result in the repository and the files that define the permissions and behavior of the assistant.

01

Generated code and dependencies

  • Access checks per record and per tenant in the generated routes
  • Commands, file access and outgoing requests in the generated code
  • Dependencies the assistant added and their known vulnerabilities
  • Changes under .github/workflows/, because workflow files can grant permissions in the repository
Claude Code writes routes, data access and configuration, and only an audit shows whether every route checks access and tenant.

02

Secrets, history and MCP

  • Keys and tokens in the code, in .env files and in earlier commits
  • Deny rules such as Read(./.env) in .claude/settings.json and whether they cover the right files
  • MCP servers in .mcp.json and their credentials; servers from other scopes are not in the repository
  • Commit trailer: Claude Code adds “Co-Authored-By” to commits by default, unless that was changed or turned off
The audit covers the entire git history, not only the current state, and the settings that are meant to block read access to secrets.

03

Permissions, instructions and pipeline

  • CLAUDE.md, .claude/rules and AGENTS.md: which instructions exist and whether the code follows them
  • Permissions in .claude/settings.json: allow and deny rules, hooks, permission mode (bypassPermissions skips prompts) and installed mods (plugins that can approve calls)
  • Sandbox: according to the documentation off by default and limited to shell commands, not file tools, MCP servers and hooks
  • Tests, build and deployment: what runs automatically and what only works locally
Scope boundary and notes
According to the documentation, Claude Code treats CLAUDE.md as context and not as enforced configuration. The client enforces permission rules and hooks regardless of what the model decides, but an installed mod can override some of them in certain setups.

Audit levels

Three audit levels, bookable separately

The audit says what has to be fixed before launch. The pentest proves what can be exploited. Continuous review accompanies changes during operation within the agreed scope.

01

Production Readiness Audit

  • Code, configuration, data access and operational readiness
  • Split: fix before launch or later
  • Action plan in order of priority
from €1,490 excl. VAT

02

Pentest

  • Attack on a test instance, with insight into the source code
  • Confirmed findings with proof, every finding with a status
from €1,000

03

AI Code Review

  • Review of changes during operation, within the agreed scope
  • You keep developing with AI tools
  • Fixed monthly price
from €990 excl. VAT / month

Process

From handover to operations

You can keep developing the application yourself with AI. We audit, fix on request and take over operations.

  1. 01

    Handover

    Repository, test instance and test accounts per role. We verify that the application belongs to you.

  2. 02

    Audit

    Audit of code and operations, on request a pentest against the test instance. Your production system stays untouched.

  3. 03

    Report

    Findings with proof, severity and remediation, plus a clear order.

  4. 04

    Fix and operate

    We fix the findings in the code and take over hosting and operations if you wish.

Frequently asked questions about auditing Claude Code apps

Answers to the most important questions

Reviews & projects

Software projects with reliable outcomes

Proof for modernization, API extension, architecture, deployment and ongoing operations.

Networking & Provisioning

WZ-IT moved our studio infrastructure from decentralised individual devices to a central platform: every site is securely connected via VPN, new devices are onboarded automatically and an entire site is provisioned from a template, without manual steps on location. What impressed me most is the breadth and depth of their knowledge: Timo and Robin are not a typical IT provider who sets up a server and leaves. The two of them think their way into highly complex infrastructure and software topics, work through every requirement we put in front of them, and build networking, provisioning and operations so that everything fits together in the end. WZ-IT is an excellent partner for complex software, network and architecture projects.
Steve KirchnerManaging Director, nextGYM GmbH
View project

International

Built in Germany's Ruhr Valley. Running worldwide.

WZ-IT designs, develops and operates infrastructure and software for clients in Germany and internationally. We deliver projects remotely and continue supporting them in ongoing operations after go-live.

Selected projects

Read client reviews

Enquiry

Have your Claude Code app audited

Describe the application, the tool and the current state. You speak directly with the people who are technically responsible.

  • You speak directly with Timo and Robin, who are technically responsible
  • A professional assessment of your situation
  • Concrete next steps for infrastructure, software or AI

The conversation is free and non-binding.

WZ-IT's advice on our Azure migration was technically sound and completely non-binding right from the intro call - we took away a great deal.
Jakob ÖschlbergerInno7 GmbH
Timo and Robin, founders of WZ-IT

What should happen next?

Choose the appropriate starting point for the current product stage.

We usually respond within one business day. Please do not send access credentials yet.

WZ-IT's advice on our Azure migration was technically sound and completely non-binding right from the intro call - we took away a great deal.
Jakob ÖschlbergerInno7 GmbH