Skip to content

Context Actions

Right-clicking a node

The menu is built from what the node actually is, in a fixed order: an "Open" action first — the same drill-in behavior described in Navigating — followed by whatever applies to this specific node:

  1. Open — always first.
  2. Rename — workspace nodes only.
  3. Session lifecycle — Resume or Pause, Restart, and Reveal Home, then Kill, for any node with a live session behind it (a session node, or an agent node with one running).
  4. Simulate Synapse Flow — agent and session nodes only. See below.
  5. Start Team… and Mesh Inbox — agent nodes only. Start Team… opens the same team-creation form as the Room's own toolbar, with this agent pre-selected as a member; Mesh Inbox opens this agent's message view — see Mail & Counters for what's inside it.
  6. A destructive action, worded for what's actually being removed — see the table below. Not every kind has one.
  7. Save Bit… — always last, on every node regardless of kind.

Destructive actions by kind

Node kindActionNotes
Bit (captured note)Delete Bit
Bit page (session)Remove Bit Page
FileRemove File
AgentDelete Agent
WorkspaceDelete WorkspaceOnly offered when more than one workspace exists
MemoryDelete Memory
SessionTerminated via Kill in the session-lifecycle group above, not a separate destructive entry
Task, Decision, Handoff, Artifact, ConflictRead-only projections of the knowledge fabric; nothing here originates from the map

Right-clicking empty canvas

A separate menu, for creating rather than acting on something that already exists: New Shell Agent, New Bit Page, New File, then, after a divider, New Agent and New Workspace. Each creates its target in the active workspace and selects it immediately.

Simulate Synapse Flow is a visualization only

Both the per-node Simulate Synapse Flow action and the toolbar's Preview Synapse Flow button (see Navigating) do the same kind of thing: they animate a cascade of pulses travelling outward along real connections, starting from a real node. Nothing about running one reads live agent state beyond the graph's existing structure, and nothing about running one triggers, messages, or otherwise acts on any agent. It is a demonstration of how activity would propagate through the shape of your current graph — never a stand-in for, or a trigger of, real agent activity. Treat anything you see ripple across the map during a simulation as exactly that: a simulation.

Where to go next

Built with purpose.