Quick question for those using multiple AI coding tools.
I use Cursor (IDE) + Claude Code (CLI) and have ~15 rules I want to keep in sync across .mdc, .claude/rules/, and AGENTS.md. Right now that means manual copying, format translation, and guessing whether my rules are too long for Cursor (<50 lines) or too heavy for Claude’s context budget.
So I’m building Rulix — an open-source CLI to write rules once and generate the right format per tool.
Follows real best practices (e.g. individual .claude/rules/ files instead of a bloated CLAUDE.md)
Side note: the documented RULE.md folder format still doesn’t work as of v2.3.23 (confirmed by forum bug reports), so Rulix generates .mdc files since that’s what actually works.
Curious:
Anyone else juggling multiple tools like this?
Do you validate rules / track budgets, or just eyeball it?
Interestingly, I was looking for something like this earlier today.
I do use multiple AI’s (OpenAI’s Codex, Claude Code, Cursor, etc.) but they are actually starting to understand each others config files - so I’m not sure that is going to be a long-term problem.
The problem I’m encountering is how to manage rules across projects. e.g., I don’t want to use my Python rules with my TypeScript projects and I want projects that use dependency X to have this rule but not that rule.
I like having the rules at the project level (not globally) and committed to version control. But that makes it difficult to use them across projects - copying and pasting is no fun. And they tend to evolve at different rates, so Project A may have better instructions than Project B.
I’m mainly looking for feedback and thoughts from the community — any input is really appreciated.