There are lots of AGENTS.md, CLAUDE.md, rules etc. While working with multiple coding agents in this kind of big project(s) it is almost best way to utilize agents from one directory.
The system is simple, create one ".agents" directory and link all ".claude" - ".opencode" and the list goes on to link that directory. In this way, there would be no drift between both normal & AI Agent coding guideline docs.
Example project: https://github.com/calcom/cal.diy
There are lots of AGENTS.md, CLAUDE.md, rules etc. While working with multiple coding agents in this kind of big project(s) it is almost best way to utilize agents from one directory.
The system is simple, create one ".agents" directory and link all ".claude" - ".opencode" and the list goes on to link that directory. In this way, there would be no drift between both normal & AI Agent coding guideline docs.
Example project: https://github.com/calcom/cal.diy