bdambrosio camera-capture
Requests one chatter/camera/capture from the ChatterBot Pi, waits for the
matching chatter/camera/image reply (filtered by request_id, since the topic
is broadcast), and saves the JPEG under the canvas server's allowlisted local
root. The frame is attached to your visual input for the rest of this turn, so
you can reason about its contents directly (no separate "describe" step) — read
it like any image you were shown.
The camera is mounted on the moving head — framing depends on where the head
points. If you need a specific view, issue head-move first and let it settle,
then capture.
The observation also contains an <img> tag and a /local URL; pass that to
the display tool (format=html) when you also want the user to see it on
screen. Only the most recent capture is kept in view.
Examples
{"thought": "user asked what I can see right now", "tool": "camera-capture"}