AI code in the system
AI Code Check
Your teams let AI write code. We check how that code gets into your system: reviews, tests, dependencies, secrets and agent permissions. Security first. The result: the biggest issues, so you know where to start.
30-minute intro callWhen it fits
- Your teams use AI tools such as Copilot, Claude Code or Cursor every day.
- Agents have access to repositories or pipelines.
- Management wants AI but fears the risk.
- Code quality has dropped since AI came in.
What we examine
- What your agents may do: rights in the repository, the pipeline and the shell
- Reviews of AI-written code
- Tests, including those AI wrote
- Checks in the pipeline
- New dependencies that AI suggested
- Secrets in code, prompts and configuration
We flag risks in data flows to AI providers and from prompt injection when we see them.
How it runs
- Intro call, 30 minutes, free: goal and scope.
- A fixed-price offer.
- Kickoff, one hour: access and contacts.
- Analysis with read access, without tying up your team.
- A written report.
- Read-out, two hours, with your team and the decision-makers.
What you receive
- The biggest gaps, ranked by urgency.
- Short rules for using AI: what agents may and may not do.
- In the same report format as the Technical Health Check, with a read-out.
What it is not
- No penetration test
- No legal advice on data protection
The facts
- Who checks
- Damir Majer. On large assessments together with a coach from ASE Academy.
- Booking
- On its own or as a module of a Technical Health Check.
- Tools
- Independent of the tool: Copilot, Claude Code, Cursor, Codex, SAP Joule and others.
- Language
- German or English
- Location
- Remote, or on site in Germany, Austria and Switzerland. Other regions by agreement.
- Technologies
- SAP ECC and S/4HANA (ABAP), Java, TypeScript and JavaScript, C#, Python
- What we need
- Read access to your Git repository. Anything else your language needs, we settle in the intro call.
- Your time
- About three hours: a one-hour kickoff and a two-hour read-out.
- Method
- Reading the code, the configuration and the pipeline: agent rights, reviews, tests, dependencies, secrets. Plus the kickoff conversation.
- Duration
- Depends on the system; agreed after the intro call.
- Price
- Fixed price after the intro call.
- What drives the price
- Size of the code base, number of systems, days on site, depth of the report.
Afterwards
- The findings are yours, whoever implements them.
- Implement them with MAJCON, or have your AI development process built with you.
- Develop your developers: at ASE Academy.
- Repeat it later, as a point of comparison.