Model picker card

AI input surface with a model chip, attach and slash command buttons, a token counter, and a keyboard hinted send button.

Saved prompt cards in a filterable grid showing title, description, category badge, and use count with a create prompt affordance.

A shared prompt panel above two side by side response cards, each with a model badge, answer body, latency and token stats, and a prefer button.

Ruled rows of recurring assistant tasks, each with cadence, delivery target, last run badge, next run time, and a pause switch, plus inline errors.

AI generated thread summary with an overview paragraph, checked decision list, owner assigned action items, and regenerate and copy controls.

Inspect grouped tool calls, commands, file reads, and their results.

Answer multiple agent questions together, with choices and optional notes.

An agent answer that combines metrics, a chart, and a results table.

Browse changed files and inspect an expandable, readable code diff.

Navigate alternative conversations and resume an earlier branch.

Compose a prompt with images, file references, and reusable skills.

Queue follow-ups while an agent works; reorder, edit, or send them now.