refactor(cli): remove dead command surface (dup dependency decorator, models typo-stub, no-op nodes refresh)

comfy_cli/cmdline.py

@@ -1435,12 +1435,6 @@ def env():
     renderer.emit(data, command="env")
 
 
-@app.command(hidden=True)
-@tracking.track_command()
-def models():
-    rprint("\n[bold red] No such command, did you mean 'comfy model' instead?[/bold red]\n")
-
-
 _FEEDBACK_DISABLED_NOTICE = (
     "[yellow]Feedback not sent — telemetry is opted out via DO_NOT_TRACK / COMFY_NO_TELEMETRY.[/yellow]\n"
     "Unset that to send, or open an issue: https://github.com/Comfy-Org/comfy-cli/issues/new/choose"
@@ -1544,9 +1538,9 @@ def agent_review(
     )
 
 
-@app.command(hidden=True)
 @app.command(
-    help="Given an existing installation of comfy core and any custom nodes, installs any needed python dependencies"
+    hidden=True,
+    help="Given an existing installation of comfy core and any custom nodes, installs any needed python dependencies",
 )
 @tracking.track_command()
 def dependency():

comfy_cli/command/nodes.py

@@ -848,25 +848,3 @@ def categories_cmd(
             renderer.console().print(tbl)
             rprint(f"[dim]{len(flat)} categories[/dim]")
     renderer.emit(payload, command="nodes categories")
-
-
-# ---------------------------------------------------------------------------
-# refresh — object_info is fetched live; nothing to cache
-# ---------------------------------------------------------------------------
-
-
-@app.command(
-    "refresh",
-    help="object_info is fetched live from the server on each command — nothing to refresh.",
-)
-@tracking.track_command("nodes")
-def refresh_cmd(
-    where: Annotated[
-        str | None,
-        typer.Option("--where", show_default=False, help="Override the resolved routing mode."),
-    ] = None,
-):
-    """Explain that object_info is fetched live and exit."""
-    renderer = get_renderer()
-    rprint("[dim]object_info is fetched live from the server on each command — nothing to refresh.[/dim]")
-    renderer.emit({"refreshed": False, "reason": "live_fetch"}, command="nodes refresh")

comfy_cli/command/run/__init__.py

@@ -531,7 +531,7 @@ def execute_cloud(
             renderer.error(
                 code="cql_no_graph",
                 message=f"could not load cloud object_info for conversion: {e}",
-                hint="run `comfy nodes refresh --where cloud` to populate the cache",
+                hint="object_info is fetched live — check your cloud sign-in and connection then retry, or run against a local server",
             )
             raise typer.Exit(code=1) from e
         try:

comfy_cli/discovery.py

@@ -48,7 +48,6 @@
     "comfy nodes path": "nodes",
     "comfy nodes types": "nodes",
     "comfy nodes categories": "nodes",
-    "comfy nodes refresh": "nodes",
     # workflow editing
     "comfy workflow slots": "workflow",
     "comfy workflow set-slot": "workflow",

comfy_cli/skills/comfy-debug/SKILL.md

@@ -47,7 +47,7 @@ comfy --json run --workflow same_workflow.json --where cloud   # just resubmit 
 If several jobs from one batch died together with this code, resubmit them all; one retry each is normally enough.
 
 ### `workflow_not_api_format`
-UI-format workflows are converted to API format **client-side** using object_info (no server conversion endpoint exists). If conversion fails with `conversion_error`, re-export via `File > Export (API)` in ComfyUI; if object_info can't be fetched (`cql_no_graph`), run `comfy nodes refresh --where cloud` or start a local server.
+UI-format workflows are converted to API format **client-side** using object_info (no server conversion endpoint exists). If conversion fails with `conversion_error`, re-export via `File > Export (API)` in ComfyUI; if object_info can't be fetched (`cql_no_graph`), retry (it's fetched live — check cloud sign-in/connection) or start a local server.
 
 ### `workflow_invalid_json`
 The file isn't valid JSON. Inspect the first/last 100 bytes — often it's an HTML error page that was saved with `.json`.

comfy_cli/skills/comfy/SKILL.md

@@ -499,7 +499,7 @@ object_info through the routing chain with a cached fallback — cloud-signed-in
 works with no local server. If the live fetch fails, the command still succeeds
 from cache and the envelope carries `data.stale: true` +
 `warnings[] {code: "object_info_stale"}` — treat results as possibly outdated
-and run `comfy nodes refresh --where cloud` to refresh.
+re-run the command once the live fetch recovers to pick up fresh object_info.
 
 Slot addresses are `<instance_id>.<input_name>`. Feed them to
 `workflow set-slot` / `workflow vary` in the Execution half. Works on