Agents and API
Your assistant can add a feedback button for you
Ask your assistant to put a feedback button on a page and it now finds the embed code right in your feedback request, with how to place and style it, instead of reaching for the share link. The developer docs gained copy-paste examples in curl, JavaScript and Python. Your private link's secret is no longer repeated in what your assistant reads, so it can't end up in a log by accident.
Agents and API
Developer docs for your assistant
Your assistant can now find everything it needs to work a link in one place: every kind of link, every read and action with its fields, and how to work a link well. Every link it reads points there.
Agents and API
Your assistant can check for only what is new
Your assistant can now ask a link for only what changed since its last look: new responses, comments and status changes. When nothing changed, it gets an empty list.
Agents and API
Assistants say who they act for
When your assistant leaves feedback, a comment or a status change for you, it now shows as "Adrian via Claude Code". The name is self-reported and not verified.
Agents and API
Your assistant can read attached files
Your assistant can now read the files people attach to responses, takes, comments and testimonials, with a link to each and who added it. Assistants read files but never upload them.
Agents and API
Every link tells your assistant how to work it
Every private link your assistant reads now carries a short note: who owns a ticket's status, the shared practices, what sends email, and what it may do on that kind of link.
Agents and API
Assistants can take part in discussions
Your assistant can now reply to a question, set a stance on a point, dismiss a question, point a question at people and add people, all under its own name.