/mind-clone-ask

description: Ask a mind-clone of an author — advice, or their likely view on a topic
argument-hint: <author> <question>  |  what would <author> think about <topic>

Invoke the **mind-clone** skill in **CHAT mode** for this request.

Arguments: $ARGUMENTS

• Resolve which clone is meant: match `$ARGUMENTS` against the folders in `./clones/`. If the author is named, use that clone; if only a question is given and exactly one clone exists, use it; if ambiguous or none exists, list `./clones/` and ask (do NOT start a build from this command).
• Then follow `reference/04-clone-runtime.md`: answer **only through the contour** (`loop.py`: `intent → sources → map → select → compile → verify`). Route by question type — how-to/method from `render playbook` (concrete steps), advice grounded in the delivered EPs, or a confidence-rated prediction (clearly labeled as extrapolation). Grounded answers quote the author with deep-links to the exact minute.
• Follow the language protocol: reply in the user's chat language (from `config.json`); keep the canon (the `ep` table) in English.
• This command never builds or modifies a clone — it only reads. To build, use `/mind-clone-build`.