Module commands 

The commands module: the command graph, projected.

Three capabilities, all read-only, all answered by crate::command_graph — the same graph the terminal renders, the same graph the workflow bridge is written from, and the same graph a shell’s completion is answered from. Declaring them here is what puts the command graph on MCP, on HTTP, in the OpenAPI document, in Swagger UI, in the Cockpit and in the generated reference, without any of those carrying a route, a schema or a sentence of its own.

Why three, and not one

Because a client that wants to know what it may call should not have to read every argument of every command to find out. commands.list answers with one line per command; commands.get answers with one command in full; commands.graph answers with the whole document and its diagnostics. Summary first, detail on demand — the shape an agent’s context needs, and the shape a command palette needs, are the same shape.

Structs

CommandFilter

Which commands to answer with. Every filter is the graph’s own vocabulary; none is a search over rendered text.

CommandGraphReport

The whole graph, with what its build found.

CommandIndex

The commands this repository offers, filtered.

CommandRequest

One command, by its canonical identity.

CommandSummary

One command, as an index shows it: enough to choose, never enough to have to skim.

Constants

COMMANDS_URI

The URI under which the index of commands is read as an MCP resource.

Functions

module

The commands module: every command this repository offers, projected once.