Interfaces¶

Open this section when the question is contractual: which agent workflows, commands, artifacts, and imports callers or operators may treat as stable instead of incidental.

Contract Model¶

flowchart LR
    caller["caller or operator"]
    surfaces["commands and APIs"]
    workflow["workflow contracts"]
    traces["trace artifacts"]
    imports["public imports"]
    consumers["runtime or automation"]

    caller --> surfaces --> workflow --> traces --> consumers
    imports --> workflow

The interface story for agent is not just about entrypoints. It is about which workflow surfaces and trace artifacts the package is prepared to defend once another tool or operator starts depending on them.

Read These First¶

open Operator Workflows first when the contract question is really about how a workflow is supposed to be run or observed
open CLI Surface when the issue begins with an agent command or entrypoint
open Compatibility Commitments when a workflow surface change may break automation or callers

Contract Risk¶

The main contract risk here is letting workflow behavior look stable to operators while its actual boundary stays undocumented.

First Proof Check¶

src/bijux_canon_agent/interfaces and apis for named caller-facing surfaces
trace-bearing artifacts and examples for workflow expectations
tests for determinism and compatibility evidence

Pages In This Section¶

Leave This Section When¶

leave for Foundation when the contract dispute is really a package-boundary dispute
leave for Architecture when a surface question reveals structural drift underneath it
leave for Operations or Quality when the boundary is clear and the question becomes execution or proof

Design Pressure¶

If a caller can only learn what is stable after reading workflow internals, the contract page is too weak. This section has to name which surfaces and trace artifacts are intentionally reusable.