Understand a new codebase — the map before the code
Day one on an unfamiliar repository is the same everywhere: hundreds of files, a README that describes an older version of the truth, and one person who knows where things are — currently on vacation. Reading files in alphabetical order is not a strategy. What you need first is not code, it's a map.
prism0x2A scans the repository on your machine — nothing is uploaded, which matters precisely when the codebase is new to you and you don't yet know what's sensitive in it — and answers the four questions every onboarding actually consists of.
Where does anything live?
The Blueprint Explorer draws the module and dependency graph of the repository as it actually is — not as the README remembers it. You see the layers, the hubs everything flows through, and the corners nothing points to, before you open a single file.
What is this codebase for?
The AMBER layer groups the code into named capabilities — discrete units of business logic. That list is the table of contents the repository never had: instead of 800 files, you start from a dozen things the system does.
What depends on what?
Every edge in the graph is a real import, extracted from the AST. Before you touch a module you can see its blast radius — who calls it, what it reaches, and whether it sits inside a cycle you'd rather know about first.
Who do I ask when nobody's around?
Ask BLUE is chat over your own scan: “where is authentication handled?”, “what writes to this table?” The answers come from the semantic model of your code — locally, using your own LLM key.
The honest part
A map does not replace reading the code — it tells you where to start, not what it means. You will still spend your first week in the editor; prism0x2A's job is to make sure that week is spent in the right files. The metrics behind the map are published — recompute them if you doubt them.