MCP: describe_command and execute_command accept id again (#365 follow-up)

#365 rejects arguments missing from a tool's schema; both tools read `id`
as an alias of `command`, so declare it.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

crates/automation/src/tests.rs

@@ -1053,3 +1053,14 @@ fn panicked_render_worker_finishes_with_an_error() {
     let reply: Value = serde_json::from_slice(&out).unwrap();
     assert_eq!(reply["result"]["isError"], true, "{reply}");
 }
+
+/// `id` is an accepted alias of `command` (#365 made unknown keys an error).
+#[test]
+fn command_tools_accept_id_as_an_alias() {
+    let mut s = server();
+    let d = call_json(&mut s, "describe_command", json!({"id": "comp.new"}));
+    assert_eq!(d["id"], "comp.new", "{d}");
+    call_json(&mut s, "execute_command", json!({"id": "comp.new", "params": {"name": "Alias", "width": 64, "height": 36}}));
+    let (_, err) = call(&mut s, "execute_command", json!({}));
+    assert!(err, "neither key is an error");
+}

crates/automation/src/tools.rs

@@ -537,7 +537,12 @@ static TOOLS: &[ToolDef] = &[
         name: "describe_command",
         description: "One command in full: id, label, menu path, shortcut, the `params` doc, a JSON Schema of its parameters (`schema`: properties, required, additionalProperties) and whether it can run now (`why` if not). Use before execute_command when unsure of the parameters; execute_command rejects unknown keys.",
         bridge_only: false,
-        schema: || schema(json!({"command": {"type": "string", "description": "Command id, e.g. `layer.newText`."}}), &["command"]),
+        schema: || {
+            schema(
+                json!({"command": {"type": "string", "description": "Command id, e.g. `layer.newText`."}, "id": {"type": "string", "description": "Alias of `command`."}}),
+                &[],
+            )
+        },
         run: describe_command,
     },
     ToolDef {
@@ -548,9 +553,10 @@ static TOOLS: &[ToolDef] = &[
             schema(
                 json!({
                     "command": {"type": "string", "description": "Command id, e.g. `comp.new`, `layer.newText`, `effect.apply` (see list_commands)."},
+                    "id": {"type": "string", "description": "Alias of `command`."},
                     "params": {"type": "object", "description": "Command parameters as documented by list_commands.", "additionalProperties": true}
                 }),
-                &["command"],
+                &[],
             )
         },
         run: execute_command,

docs/agents.md

@@ -80,8 +80,8 @@ See [MCP conventions](mcp.md) for core tools, resources, strict arguments, and r
 | Tool | What it does |
 |---|---|
 | `list_commands {filter?, enabled_only?, schemas?}` | Discover command ids, their param docs (and, with `schemas`, a JSON Schema each), and whether each can run now. |
-| `describe_command {command}` | One command in full: label, menu, shortcut, params doc, JSON `schema`, enabled / `why`. |
-| `execute_command {command, params?}` | Run any command (undoable). Unknown parameter keys are rejected with the accepted list. |
+| `describe_command {command \| id}` | One command in full: label, menu, shortcut, params doc, JSON `schema`, enabled / `why`. |
+| `execute_command {command \| id, params?}` | Run any command (undoable). Unknown parameter keys are rejected with the accepted list. |
 | `get_state` | Editor state (active comp, time, selections, tool) plus `app`: version, command and effect counts, export formats, parity summary. |
 | `run_script {code, name?}` | Run JavaScript with the After Effects-style scripting object model (`app.project`, `comp.layers.addText(…)`, `layer.property("ADBE Transform Group").property("ADBE Position").setValueAtTime(…)`…). Returns `{ok, result, output, error: {message, line, column}}`; edits are undoable. |
 | `get_project` / `get_comp {comp?}` | Project items, comp settings and layers. |