Where my hug at?

What you remember of Hugging Face is as old as your training and misses the long tail. So don't pick from memory: Where My Hug At searches Hugging Face for each part of the need, you pick from what it finds, and it checks every pick live and writes the report the user sees. Three steps, always in order.

With the MCP tools

If the where_my_hug, where_my_hug_search and where_my_hug_verify tools are connected, call where_my_hug with the need and follow what it returns. It says what to ask the user first, what to search, and when to call where_my_hug_verify.

Without them, from a shell

The same three steps run from a shell. Node 22.18 or later is needed. Write the JSON to a file and pass it on stdin:

1. npx -y where-my-hug rubric < details.json, where the file holds {"need": "...", "compute": "...", "license": "...", "stack": "...", "data": "..."}. Fill what the conversation or codebase already says and leave the rest out. If it returns questions, ask the user, then run it again with the answers ("unknown" is fine).
2. Follow the method it returns. Where it says to call where_my_hug_search, write {"parts": [...]} to a file and run npx -y where-my-hug search < parts.json.
3. Pick from what it found. Where the method says to call where_my_hug_verify, write that JSON to a file and run npx -y where-my-hug verify < slate.json. If its JSON says SEARCH_FIRST or RECALL_REPLACEMENTS, do what its instructions say and run it again.

Show the user the markdown report as it comes back, without restating its facts from memory. Public models and datasets need no token. Private or gated ones need the user's Hugging Face token in the server's own configuration (the Claude Code plugin asks for it); never ask for one in the chat.