Skip to main content

Using AI

Camunda follows an AI-first approach. Follow the company AI policies and apply the guidelines below when using AI for frontend work.

Your responsibility​

AI writes code, you own the result. Your job is to steer agents toward a good solution, not accept the first thing that compiles. Apply critical thinking, iterate, and refine.

A PR should look as if AI was not involved. If it doesn't, you are shifting review burden to other engineers and wasting cycles from automated agents reviewing your PRs.

Working with agents​

  • Use the frontend agent for frontend work. It is purpose-built for the orchestration-cluster webapp and knows our conventions. The agent is currently being developed and will be available soon.
  • Break the problem down before prompting. Smaller, well-scoped prompts produce better results than "implement this feature." Use plan mode to get the agent's help with the breakdown.
  • Review each response critically. Check for unnecessary abstractions, dead code, wrong patterns, and deviations from project conventions.
  • Iterate. Push back on the agent when the solution is not good enough. Ask it to simplify, restructure, or try a different approach.
  • Use the project docs (this site) as context for the agent so it follows our conventions.

Before opening a PR​

  • Code follows project conventions (see sibling docs).
  • No leftover boilerplate, redundant comments, or over-engineered abstractions typical of AI output.
  • You can explain what your code does.
  • Tests are meaningful, not just "make the build pass."
  • The diff is clean and reviewable, same standard as hand-written code.