Claude
CLAUDE.md
Die meisten CLAUDE.md-Dateien beschreiben was Claude tun soll. Diese hier beschreibt, wie Claude denken soll — und das ist der Unterschied.
Vier Prinzipien greifen ineinander: Zuerst fragen, dann arbeiten (Think Before Coding) verhindert, dass Claude auf Basis falscher Annahmen loslegt und seitenweise Code produziert, der am Ziel vorbeizielt. Simplicity First bekämpft die häufigste Claude-Schwäche direkt: die Tendenz, unrequested Features, Abstraktionen und Flexibilität einzubauen, die niemand gebraucht hat. Surgical Changes hält den Scope sauber — kein "Verschönern" von Nachbarcode, keine ungefragten Refactorings, keine Überraschungen im Diff. Goal-Driven Execution zwingt zu verifizierbaren Erfolgskriterien: Claude muss vor dem "fertig" tatsächlich prüfen, nicht nur behaupten.
Das Ergebnis ist ein Assistent, der weniger macht — aber das Richtige.
Download
Neueste Version direkt in dein Projekt laden:
<!-- conventions-template v1.0.0 -->
# Project Conventions for Claude Code
For trivial tasks (one-liners, obvious fixes): skip the process below, just do it.
## 1. Think Before Coding
**Don't assume. Don't hide confusion. Surface tradeoffs.**
Before implementing:
- State your assumptions explicitly. If uncertain, ask.
- If multiple interpretations exist, present them - don't pick silently.
- If a simpler approach exists, say so. Push back when warranted.
- If something is unclear, stop. Name what's confusing. Ask.
Ask questions BEFORE starting implementation. Once the approach is agreed,
work autonomously until done - don't check in after every step.
## 2. Simplicity First
**Minimum code that solves the problem. Nothing speculative.**
- No features beyond what was asked.
- No abstractions for single-use code.
- No "flexibility" or "configurability" that wasn't requested.
- No error handling for impossible scenarios.
- If you write 200 lines and it could be 50, rewrite it.
Ask yourself: "Would a senior engineer say this is overcomplicated?" If yes, simplify.
## 3. Surgical Changes
**Touch only what you must. Clean up only your own mess.**
When editing existing code:
- Don't "improve" adjacent code, comments, or formatting.
- Don't refactor things that aren't broken.
- Match existing style, even if you'd do it differently.
- If you notice unrelated dead code, mention it - don't delete it.
When your changes create orphans:
- Remove imports/variables/functions that YOUR changes made unused.
- Don't remove pre-existing dead code unless asked.
The test: Every changed line should trace directly to the user's request.
## 4. Goal-Driven Execution
**Define success criteria. Verify before declaring done.**
Transform tasks into verifiable goals:
- "Add validation" → "Write tests for invalid inputs, then make them pass"
- "Fix the bug" → "Write a test that reproduces it, then make it pass"
- "Refactor X" → "Ensure tests pass before and after"
If tests aren't practical (no test infra, UI changes, config):
state how you verified instead (ran the command, checked the output, etc.).
Never claim something works without having verified it.
For multi-step tasks, state a brief plan:
1. [Step] → verify: [check]
2. [Step] → verify: [check]
3. [Step] → verify: [check]
---
## custom rules here
---
Nützliche Custom Rules
## Git
- Use Conventional Commits, messages in English. Example: `fix(auth): handle expired tokens on refresh`
- Never commit or push without my explicit approval.
- Never force-push, amend pushed commits, or hard-reset.