Node View
Node View is the architecture of your paranet, drawn from the source of your node or the local development package you are working with. It answers what exists, what each actor can do, and how they depend on one another.
Read it left to right: skillsets and skills on the left, the actors that implement them, and the dependencies between them.
What you will find
- Actors — the participants on your paranet
- Skills — a capability, addressed as
subject/action - Skillsets — a named group of skills
- Contextual labels — the values used to match a request to a fulfiller
- Observers — actors watching conversations they are not party to
Select any entity on the canvas to open its details panel. Use the left panel to search by name, or to jump straight to an actor or skillset.
Actor details
An actor's panel covers three things:
What it provides — the skills it implements, grouped by skillset, and the skills it exposes through external implementations.
What it depends on — every skill the actor requests from elsewhere, and which actor on this paranet provides it. A dependency provided by nothing is called out, which is worth knowing before it fails at runtime rather than after.
What breaks without it — select Show impact to see which actors stop working if this one does. The calculation follows dependencies transitively and accounts for redundancy: an actor is only affected when a dependency loses every provider it had.
Skill details
A skill's panel is its contract:
- Input and output — the fields it accepts and returns. Required fields are marked, and each carries the description from its schema.
- Constraints and labels — the matching rules that decide which actor fulfils a request. Shown when the skill declares them.
- Questions — if the skill asks the requester anything mid-conversation, each question appears alongside the answer it expects.
- Capabilities — badges indicating that the skill ships an input form or an Adaptive Card for its output.
Only the sections a skill actually declares are shown, so a panel with few sections describes a simple skill rather than a missing one.
Usage
The architecture describes what your paranet can do. Switch on Usage, top right of the canvas, to see what it actually does.
Paracord reads recent conversations from the Ledger and marks the diagram:
| Mark | Meaning |
|---|---|
| Thicker edge | more calls along that dependency |
| Faded, dashed edge | declared, but never called |
| Moving edge | called within the last 15 minutes |
| Faded node | a skill nothing has called |
| Filled node | a skill with failures |
Hover a skill for its call count, or open its panel for the breakdown: how many completed, how many failed, how many calls arrived directly versus from inside another conversation, and the message volume across them.
That last distinction is useful when reviewing a design. A skill only ever called from within another conversation is internal machinery; a skill called directly is an entry point.
Usage is read from a recent window of conversations rather than the entire Ledger, and the canvas states how many it scanned. Treat "never called" as "not called within that window".
Where to go next
- Netflow shows these same actors executing, live.
- The Ledger holds every conversation between them, permanently.
- Actor Hub lets you call any skill on this canvas yourself.