Claude Code
Claude Code is a coding tool from Anthropic that works with the files and tools of a software project. Developers can describe a bug fix, ask for an explanation of unfamiliar code or request a change across several files. Claude Code can inspect the repository, edit code and run commands, connecting instructions in ordinary language with the project's development process.
For a company in Frankfurt, Claude Code brings this help to the work that teams repeat every week, such as reports, data exports and document checks. The team reviews each result before it uses it, and access permissions and the choice of material supplied to the tool remain part of the controls of the company.
What Claude Code is, and what it is not
Claude Code maps and explains a codebase, reads issues and writes code, runs tests, opens pull requests and follows imports across repositories, and it keeps working when something breaks. You steer it from a terminal, from VS Code and Cursor, from JetBrains IDEs, from Slack, from the web, from the Claude desktop app, from iOS and Android, and from GitHub Actions, on macOS, Linux and Windows. Claude Code asks for permission before it changes a file or runs a command.
Claude Code differs from the Claude chat apps, where you paste text and get text back, and it is a different thing from an autocomplete extension, which predicts the next few lines without understanding the task. Claude Code works on complete tasks that can involve multiple lines of code. On Team and Enterprise plans the Projects feature supervises several agents at once, so one person can have three tasks in flight.
What it takes over in the daily work of a company
Most of the work Claude Code takes over in a company sits outside the software department. A sales team exports its open deals from the CRM every Monday and copies them into the same Excel template by hand. Claude Code reads the export, writes a script that builds the report and runs that script again next week with one command.
The same pattern fits a contract check against the rules list of the company, a merge of CSV exports from two ERP systems or a summary of public documents. Each of these jobs repeats and follows rules that someone can write down, and that is the kind of work the agent suits.
What it does not take over is the decision. A person still reads the result and approves the change, and that review step belongs in the process of the team before the agent works with live data.
Skills, agents and hooks
A repository can carry its own instructions file, which is where a team writes down the conventions, the commands and the rules for what the agent may touch. Skills package a procedure the team repeats so the agent follows the house method instead of inventing one. Subagents split a large task so one part of the work runs separately from the main thread. Hooks run a command of yours at defined points, which is how a team enforces a check the agent cannot skip.
For a finance team those four are the control surface: the instructions file sets the boundary, the hooks enforce it, and both live in version control where a reviewer can see who changed them.
OpenAI Codex, the other tool in this category
OpenAI Codex is the comparable product from OpenAI, with a CLI, an IDE extension, a cloud service and the ChatGPT apps as its surfaces, plus sandboxing, agent approvals, auto-review and profile-based permissions. The two tools are built on the same idea and differ in the surfaces they offer and the platforms they assume.
ChatGPT and Claude compared for finance goes through the comparison, including the coding one, and explains why a platform team usually decides on the cloud and the identity provider it already runs.
What does a company settle before its teams use it?
The first question is where the data goes. The tool runs on the laptop of the employee and talks to the model API without a backend server of its own, and Anthropic offers self-hosted environments that run inside a customer network next to internal services.
Then the company decides which folders and systems the agent may touch, under which account it works and who reviews its results. The permission prompt is the mechanism, but the policy has to exist on paper first, because an employee who clicks through every prompt has disabled the control.
The EU AI Act bears on the output and not on the tool: a script that ends up in a system for hiring, credit or customer decisions falls under the obligations attached to that system, whoever wrote it.
Where do Frankfurt developers using it meet?
Developers meet at the German-language Claude meetup that the Agentic AI Community Frankfurt runs every other week at byte5 GmbH in the Speicherstraße, where the subjects include programming and web development alongside prompting and AI workflows. Claude meetups in Frankfurt describes the format and the room.
Finance Loop runs Claude Hacker House for people who work in finance, and the CCAF certification is the route for someone who wants the architecture side on paper. Blockchain software development covers the other developer field this site follows.
Claude Code and Finance Loop
Finance Loop brings together the people in Frankfurt companies who run coding agents on their own data and workflows, and that is where the practical answers come from: what the permission policy ended up saying, which systems stayed off limits and how the review step was documented. Finance Loop is the meeting place for that exchange.
Finance Loop is a professional network supporting the institutional use of emerging technologies in finance, such as AI, tokenization, stablecoins, and DeFi. Finance Loop helps its members build skills and personal networks in these fields: Investment & Digital Assets, Payments & Digital Money, Digital Infrastructure & Sovereignty, and Risk & Compliance.