The rationale is simple - I want my project repos to be self-sufficient in the sense that it includes both its source code as well as goals/issues/docs. I also need a simpler/easier integration with AI agents that do a lot of work for me today. Ideally it would be sufficient to launch a Claude Code or whatever in the project dir regardless of where and what machine I clone it on or directly on GitHub and it would have the entire context:
project documentation
current issues or todos
backlog
roadmap
agent memories/skills
general notes
The requirement is also that the tool should not be an invasive species. It should not require to carry along any databases/indices/configuration of its own - nothing you wouldn't have without it
source & further reading
dev.to ā original article
Generative Engine Optimization Is Growing, but the Panda Parallel Is Not Proven
Apple quietly shipped everything you need to build a real-time translator ā so I built one
The EU AI Act asks AI to identify itself. I checked 30 years of that experiment on my server.