memvid/claude-brain — reverse-engineered prompt
Reverse engineered prompt
Build me a Claude Code plugin that gives Claude a simple local memory across coding sessions.
I want it to save important session context, decisions, bugs we fixed, and solutions into one portable .mv2 file inside .claude/mind.mv2. No cloud, no database, no API keys. It should automatically bring useful recent memory back in when a new Claude Code session starts, so I can ask things like “what did we decide about auth” and get a useful answer instead of starting over.
Please include Claude commands like /mind stats, /mind search "authentication", /mind ask "why did we choose JWT", and /mind recent. Natural wording should work too, like “search my memory for auth bugs”.
Make it fast, private, and easy to share by copying or committing the single memory file. Include a simple optional CLI for direct access to the memory file. Look up current Claude Code plugin docs online if you need to.
Want more depth? Deep Reverse