Railly/agent-brain ? reverse-engineered prompt
Reverse engineered prompt
Build me a local first personal knowledge system for coding agents, centered on Obsidian and plain Markdown. I want it to help me keep context about my stack, projects, goals, recurring mistakes, and daily work so Claude Code and Codex can pick up where I left off without a cold start.
Set it up with a simple vault structure, daily journaling, inbox capture, note linking, search, reflection, and shipping workflows. It should feel like a reusable template I can clone, initialize, and customize for my own identity and workspace. Make sure the agent commands and workflow files stay in sync, and include a way to verify that the generated pieces do not drift.
I also want a clean migration path from an older version, plus clear docs for setup and the available commands. If you need to look up current docs online while building, do that.