Omniscio

Plugin UI Contributions

An opt-in plugin contribution: a plugin can add its own buttons to the header toolbar and a session's right-click menu, and navigate the app to a session, project, inbox or virtual project. The buttons are read-only chrome until a click lands back in that plugin's own pane.

On this page

What it is

Status: shipped (always on, landed 2026-08-06). A plugin can add its own buttons to the header toolbar and the session right-click menu, and navigate the app to a session, hub, or view — there is no Lab flag for it and nothing to enable.

A plugin that opts in gets a read-only presence in the Omniscio shell. Through its webview pane it can do two things, both mediated by the host:

  1. Navigate the app — navigation.goTo({ kind, id? }) switches the active view to a core session, project, inbox, or virtual project — the same chokepoints the app itself uses when you click around. It is reversible (it never destroys anything), so it is the most benign primitive.
  2. Contribute chrome — toolbar.setItems([...]) and contextMenu.setSessionItems([...]) set the buttons the plugin wants in the header toolbar and in the right-click menu of any session. Both are replace-semantics per plugin: whatever the plugin last sent is what shows, and sending an empty list removes its items.

The contributed buttons and menu items are read-only chrome until a click lands back in the plugin's own pane — clicking one sends a command:<id> event to that plugin's webview and nothing else. Anything that actually changes data still goes through the plugin's normal approval flow.

Where to find it

There is nothing to switch on. Look at the app's own chrome: a contributing plugin's buttons appear on the header toolbar, namespaced to that plugin, and its items appear in the right-click menu of any session. The plugin's own pane is where the result of a click shows up, and it opens from wherever that plugin sits in the app.

A plugin only gets here if you consented to the navigation and/or chrome permission when you installed it, from the Plugin Marketplace.

How it behaves

  1. Nothing to enable — the feature ships always-on (it landed with the plugin-platform parity slice on 2026-08-06).
  2. A plugin opts in by declaring navigation and/or chrome in its manifest permissions list. The marketplace asks you to consent when a plugin requests a permission; a plugin that doesn't declare it can't use it.
  3. Once installed, a contributing plugin's buttons appear (namespaced plugin:<pluginId>:<id>) on the header toolbar, and its items appear in the right-click menu of any session.
  4. Click one — a command:<id> event (plus the session id, for menu items) is delivered to that plugin's own webview. Open its pane to see the result.
  5. Desktop only: plugin navigation and chrome clicks are refused over the phone/WS bridge (see the implementation facts below). A listed plugin whose code ships with the app can open its own screen on a paired phone, but its toolbar buttons and menu items still do not (see mobile-plugin-ui-bridge.md).

Plugin Marketplace covers installing plugins and the permission-consent UX this page's opt-in rides on. For the higher-power capabilities a plugin can call, see Plugin Bridge Capabilities; for the in-plugin session experience — sessions started from inside a plugin's own surface — see Plugin AI-Native Sessions. The read-only disclosure that a plugin's AI can take actions is Plugin Capability Card, and how a session discovers and drives installed plugins is Plugin CLI Discovery.

Last verified 2026-09-23