fix: rename /model slash command to /models

packages/coding-agent/src/config/model-registry.ts

@@ -144,7 +144,7 @@ interface ProviderOverride {
  * Without explicit transport propagation, an openrouter (or any) entry
  * marked `transport: pi-native` in models.yml silently reverts to the
  * default openai-completions transport after the background catalog
- * refresh — so the first `/model` switch after boot hits the raw OpenAI
+ * refresh — so the first `/models` switch after boot hits the raw OpenAI
  * chat-completions URL instead of the gateway's `/v1/pi/stream` (#2555).
  * See `xiaomi-tp-discovery-merge.test.ts` and the `refresh()` baseUrl-override
  * regression in `model-registry.test.ts`.

packages/coding-agent/src/config/model-resolver.ts

@@ -1399,7 +1399,7 @@ export async function resolveAllowedModels(
 /**
  * Synchronous subset of {@link resolveAllowedModels} for contexts where async is unavailable
  * (e.g. `getAvailableModels()` which is called from the ACP model-list advertisement, RPC
- * `get_available_models`, and the `/model` slash command). Uses the same effective
+ * `get_available_models`, and the `/models` slash command). Uses the same effective
  * `enabledModels` scope semantics as startup resolution:
  *
  * - Glob selectors match `provider/modelId` and bare model id

packages/coding-agent/src/extensibility/extensions/model-api.ts

@@ -15,7 +15,7 @@ import type { ExtensionModelQuery } from "./types";
 
 /**
  * Build the `ctx.models` facade. `getModel` is read lazily so `current()` always
- * reflects the live session model (it can change mid-session via `/model`).
+ * reflects the live session model (it can change mid-session via `/models`).
  */
 export function createExtensionModelQuery(
 	modelRegistry: ModelRegistry,

packages/coding-agent/src/modes/components/advisor-config.ts

@@ -10,7 +10,7 @@
  *
  * Each screen is backed by a proven primitive — {@link SelectList} (list / detail
  * / tools / thinking), {@link Input} (name), {@link ModelSelectorComponent} (the
- * same rich `/model` picker, in direct-select mode), and {@link HookEditorComponent}
+ * same rich `/models` picker, in direct-select mode), and {@link HookEditorComponent}
  * (multiline instructions; Ctrl+G opens `$EDITOR`). The overlay edits an in-memory
  * {@link WatchdogConfigDoc} and only touches disk + the live advisors via the host
  * `save` callback.

packages/coding-agent/src/modes/components/settings-selector.ts

@@ -367,7 +367,7 @@ export interface SettingsRuntimeContext {
 	thinkingLevel: ThinkingLevel | undefined;
 	/** Available themes */
 	availableThemes: string[];
-	/** Provider/source ids shown in /model. */
+	/** Provider/source ids shown in /models. */
 	providers: string[];
 	/** Working directory for plugins tab */
 	cwd: string;

packages/coding-agent/src/modes/controllers/selector-controller.ts

@@ -694,7 +694,7 @@ export class SelectorController {
 						this.ctx.updateEditorBorderColor();
 						const roleSelectorHint = this.ctx.keybindings.getKeys("app.model.select")[0] ?? "Alt+M";
 						this.ctx.showStatus(
-							`Session-only model: ${selector ?? model.id}. Use ${roleSelectorHint} or /model for roles.`,
+							`Session-only model: ${selector ?? model.id}. Use ${roleSelectorHint} or /models for roles.`,
 						);
 						done();
 					} catch (error) {

packages/coding-agent/src/sdk.ts

@@ -2140,7 +2140,7 @@ export async function createAgentSession(options: CreateAgentSessionOptions = {}
 				modelFallbackMessage =
 					patterns && patterns.length > 0
 						? `No model available matching enabledModels (${patterns.join(", ")}) with usable credentials. Configure auth for an allowed provider or adjust enabledModels.`
-						: "No models available. Use /login or set an API key environment variable. Then use /model to select a model.";
+						: "No models available. Use /login or set an API key environment variable. Then use /models to select a model.";
 			}
 		}
 

packages/coding-agent/src/session/agent-session.ts

@@ -7849,7 +7849,7 @@ export class AgentSession {
 				throw new Error(
 					"No model selected.\n\n" +
 						`Use /login, set an API key environment variable, or create ${getAgentDbPath()}\n\n` +
-						"Then use /model to select a model.",
+						"Then use /models to select a model.",
 				);
 			}
 
@@ -9170,7 +9170,7 @@ export class AgentSession {
 
 		// Trust the recorded role only while its resolved model still IS the
 		// active model. A model switch through another surface (alt+m, retry
-		// fallback, /model) or a role re-configuration leaves the recorded role
+		// fallback, /models) or a role re-configuration leaves the recorded role
 		// pointing at a model the session no longer runs; cycling from that
 		// stale slot lands on the wrong neighbor and reads as a skipped entry.
 		const lastRole = this.sessionManager.getLastModelChangeRole();
@@ -10312,7 +10312,7 @@ export class AgentSession {
 						return stream.result();
 					},
 					telemetry: resolveTelemetry(this.agent.telemetry, this.sessionId),
-					// Honor the user's /model thinking selection on the handoff path.
+					// Honor the user's /models thinking selection on the handoff path.
 					// Clamped per-model inside generateHandoffFromContext via
 					// resolveCompactionEffort so unsupported-effort models don't trip
 					// requireSupportedEffort.
@@ -12219,7 +12219,7 @@ export class AgentSession {
 						metadata: this.agent.metadataForProvider(candidate.provider),
 						convertToLlm: messages => this.#convertToLlmForSideRequest(messages),
 						telemetry,
-						// Honor the user's /model thinking selection (incl. `off`) on
+						// Honor the user's /models thinking selection (incl. `off`) on
 						// the manual `/compact` path. Clamped per-model inside compact()
 						// via resolveCompactionEffort so unsupported-effort models
 						// (xai-oauth/grok-build) don't trip requireSupportedEffort.
@@ -12926,7 +12926,7 @@ export class AgentSession {
 									initiatorOverride: "agent",
 									convertToLlm: messages => this.#convertToLlmForSideRequest(messages),
 									telemetry,
-									// Honor the user's /model thinking selection on the
+									// Honor the user's /models thinking selection on the
 									// auto-compaction path — the most-fired compaction
 									// site. Clamped per-model inside compact() via
 									// resolveCompactionEffort.

packages/coding-agent/src/slash-commands/acp-builtins.ts

@@ -8,7 +8,7 @@ export type { AcpBuiltinSlashCommandResult } from "./types";
 /**
  * All names (primary + aliases) that are reserved by ACP builtins. Used to
  * filter out extension commands that would shadow a builtin or its alias at
- * dispatch time (e.g. `models` is an alias for `/model`, so an extension
+ * dispatch time (e.g. `/models` is the canonical builtin name, so an extension
  * registering `models` would appear in the palette but execute the builtin).
  */
 export const ACP_BUILTIN_RESERVED_NAMES: ReadonlySet<string> = new Set(
@@ -19,8 +19,8 @@ export const ACP_BUILTIN_RESERVED_NAMES: ReadonlySet<string> = new Set(
  * Whether an extension command named `name` would be captured by ACP builtin
  * dispatch before reaching the extension handler. Beyond exact name/alias
  * collisions, `parseSlashCommand` treats `:` as a name/args separator, so a
- * colon-namespaced name whose prefix is a handled builtin (e.g. `model:foo`)
- * executes the `/model` builtin with `foo` as args. Such names must not be
+ * colon-namespaced name whose prefix is a handled builtin (e.g. `models:foo`)
+ * executes the `/models` builtin with `foo` as args. Such names must not be
  * advertised to ACP clients.
  */
 export function isAcpBuiltinShadowedName(name: string): boolean {

packages/coding-agent/src/slash-commands/builtin-registry.ts

@@ -336,8 +336,7 @@ const BUILTIN_SLASH_COMMAND_REGISTRY: ReadonlyArray<SlashCommandSpec> = [
 		},
 	},
 	{
-		name: "model",
-		aliases: ["models"],
+		name: "models",
 		description: "Switch model for this session",
 		acpDescription: "Show current model selection",
 		getTuiAutocompleteDescription: runtime => {
@@ -353,7 +352,7 @@ const BUILTIN_SLASH_COMMAND_REGISTRY: ReadonlyArray<SlashCommandSpec> = [
 				);
 				if (!match) {
 					return usage(
-						`Unknown model: ${modelId}. Use ACP \`session/setModel\` for picker-driven selection or list available models with /model.`,
+						`Unknown model: ${modelId}. Use ACP \`session/setModel\` for picker-driven selection or list available models with /models.`,
 						runtime,
 					);
 				}

packages/coding-agent/test/acp-agent.test.ts

@@ -1570,13 +1570,13 @@ describe("ACP agent", () => {
 		];
 		// Extension runner: unique command + one colliding with an ACP builtin
 		// ("fast") + a colon-namespaced one whose prefix is a builtin
-		// ("model:foo" parses as builtin `/model` with args `foo` at dispatch).
+		// ("models:foo" parses as builtin `/models` with args `foo` at dispatch).
 		(session as unknown as { extensionRunner: unknown }).extensionRunner = {
 			getRegisteredCommands(reserved?: Set<string>) {
 				return [
 					{ name: "my-ext-cmd", description: "Extension command", handler: async () => {} },
 					{ name: "fast", description: "Would shadow builtin", handler: async () => {} },
-					{ name: "model:foo", description: "Colon-shadowed by /model", handler: async () => {} },
+					{ name: "models:foo", description: "Colon-shadowed by /models", handler: async () => {} },
 				].filter(cmd => !reserved?.has(cmd.name));
 			},
 		};
@@ -1611,8 +1611,8 @@ describe("ACP agent", () => {
 		// ACP builtin "fast" appears exactly once (reserved-set exclusion, no duplicate from extension).
 		expect(names.filter(n => n === "fast").length).toBe(1);
 		// Colon-namespaced collision with a builtin prefix is not advertised:
-		// ACP would dispatch `/model:foo` to the `/model` builtin, not the extension.
-		expect(names).not.toContain("model:foo");
+		// ACP would dispatch `/models:foo` to the `/models` builtin, not the extension.
+		expect(names).not.toContain("models:foo");
 
 		harness.abortController.abort();
 		await Bun.sleep(0);

packages/coding-agent/test/acp-builtins.test.ts

@@ -414,36 +414,36 @@ describe("ACP builtin slash commands", () => {
 		expect(output[0]).toContain("No messages");
 	});
 
-	// /model
-	it("model: returns current model when set", async () => {
+	// /models
+	it("models: returns current model when set", async () => {
 		const { output, runtime } = createRuntime();
 		runtime.session.model = { provider: "anthropic", id: "claude-opus-4-5" } as never;
 
-		const result = await executeAcpBuiltinSlashCommand("/model", runtime);
+		const result = await executeAcpBuiltinSlashCommand("/models", runtime);
 
 		expect(result).toEqual({ consumed: true });
 		expect(output[0]).toContain("anthropic/claude-opus-4-5");
 	});
 
-	it("model: returns no-selection message when undefined", async () => {
+	it("models: returns no-selection message when undefined", async () => {
 		const { output, runtime } = createRuntime();
 
-		const result = await executeAcpBuiltinSlashCommand("/model", runtime);
+		const result = await executeAcpBuiltinSlashCommand("/models", runtime);
 
 		expect(result).toEqual({ consumed: true });
 		expect(output[0]).toContain("No model");
 	});
 
-	it("model: returns ACP usage message when args provided", async () => {
+	it("models: returns ACP usage message when args provided", async () => {
 		const { output, runtime } = createRuntime();
 
-		const result = await executeAcpBuiltinSlashCommand("/model claude-3-5-sonnet", runtime);
+		const result = await executeAcpBuiltinSlashCommand("/models claude-3-5-sonnet", runtime);
 
 		expect(result).toEqual({ consumed: true });
 		expect(output[0]?.toLowerCase()).toContain("acp");
 	});
 
-	it("model: applies known id and emits both title + config change notifications", async () => {
+	it("models: applies known id and emits both title + config change notifications", async () => {
 		const { output, runtime, session } = createRuntime();
 		const available = [{ provider: "anthropic", id: "claude-3-5-sonnet", contextWindow: 200_000 }];
 		session.getAvailableModels = () => available;
@@ -457,7 +457,7 @@ describe("ACP builtin slash commands", () => {
 		};
 		const setModelSpy = spyOn(session, "setModel").mockResolvedValue(undefined);
 
-		const result = await executeAcpBuiltinSlashCommand("/model claude-3-5-sonnet", runtime);
+		const result = await executeAcpBuiltinSlashCommand("/models claude-3-5-sonnet", runtime);
 
 		expect(result).toEqual({ consumed: true });
 		expect(setModelSpy).toHaveBeenCalledWith(available[0]);
@@ -466,14 +466,14 @@ describe("ACP builtin slash commands", () => {
 		expect(configNotified).toBe(1);
 	});
 
-	it("model: does not emit config change when id is unknown", async () => {
+	it("models: does not emit config change when id is unknown", async () => {
 		const { runtime } = createRuntime();
 		let configNotified = 0;
 		runtime.notifyConfigChanged = () => {
 			configNotified++;
 		};
 
-		await executeAcpBuiltinSlashCommand("/model nonexistent", runtime);
+		await executeAcpBuiltinSlashCommand("/models nonexistent", runtime);
 
 		expect(configNotified).toBe(0);
 	});
@@ -1035,23 +1035,23 @@ describe("wave 5 — adapters and polish", () => {
 		}
 	});
 
-	// /model with unknown id
-	it("/model gpt-fake-9000: returns unknown-model message", async () => {
+	// /models with unknown id
+	it("/models gpt-fake-9000: returns unknown-model message", async () => {
 		const { output, runtime } = createRuntime();
-		const result = await executeAcpBuiltinSlashCommand("/model gpt-fake-9000", runtime);
+		const result = await executeAcpBuiltinSlashCommand("/models gpt-fake-9000", runtime);
 		expect(result).toEqual({ consumed: true });
 		expect(output[0]).toContain("Unknown model");
 	});
 
-	// /model with known id (fake registry)
-	it("/model known-id: reports model set and triggers notifyTitleChanged", async () => {
+	// /models with known id (fake registry)
+	it("/models known-id: reports model set and triggers notifyTitleChanged", async () => {
 		const { output, session, runtime } = createRuntime();
 		session.getAvailableModels = () => [{ provider: "anthropic", id: "claude-sonnet-test" }];
 		let titleChanged = false;
 		runtime.notifyTitleChanged = () => {
 			titleChanged = true;
 		};
-		const result = await executeAcpBuiltinSlashCommand("/model claude-sonnet-test", runtime);
+		const result = await executeAcpBuiltinSlashCommand("/models claude-sonnet-test", runtime);
 		expect(result).toEqual({ consumed: true });
 		expect(output[0]).toContain("Model set to anthropic/claude-sonnet-test.");
 		expect(titleChanged).toBe(true);

packages/coding-agent/test/agent-session-openai-completions-model-switch.test.ts

@@ -12,7 +12,7 @@ import { TempDir } from "@oh-my-pi/pi-utils";
 // Regression: `#closeProviderSessionsForModelSwitch` historically only handled
 // the `openai-codex-responses` / `openai-responses` keys and left
 // `openai-completions:<provider>:<baseUrl>:<modelId>` entries behind on a
-// /model switch. The cached strict-tools disable scopes and reasoning-effort
+// /models switch. The cached strict-tools disable scopes and reasoning-effort
 // fallbacks for the old backend then survived indefinitely — repro reported
 // in #3260 (PR #3236).
 

packages/coding-agent/test/agent-session-role-thinking.test.ts

@@ -633,7 +633,7 @@ describe("AgentSession role model thinking behavior", () => {
 		// Record a model_change for the "slow" role WITHOUT switching the
 		// active model — the session still runs the default model. This is the
 		// stale state left behind when the model is changed through another
-		// surface (alt+m, temporary model, /model) after a role cycle.
+		// surface (alt+m, temporary model, /models) after a role cycle.
 		session.sessionManager.appendModelChange(`${slowModel.provider}/${slowModel.id}`, "slow");
 		expect(session.sessionManager.getLastModelChangeRole()).toBe("slow");
 		expect(session.model?.id).toBe(defaultModel.id);

packages/coding-agent/test/interactive-mode-prompt-template-autocomplete.test.ts

@@ -201,7 +201,7 @@ describe("InteractiveMode prompt-template autocomplete (#2462)", () => {
 		expect(matches.filter(name => name === "exit")).toHaveLength(1);
 	});
 
-	it("does not duplicate templates whose names collide with builtin slash command aliases", async () => {
+	it("does not duplicate templates whose names collide with builtin slash command names", async () => {
 		const created = createHarness([
 			{
 				name: "models",
@@ -217,7 +217,7 @@ describe("InteractiveMode prompt-template autocomplete (#2462)", () => {
 		const provider = slot.current;
 		expect(provider).toBeDefined();
 		const matches = await fetchSlashSuggestions(provider!, "/models");
-		// Builtin `/model` owns the `/models` alias. The colliding template is filtered
+		// Builtin `/models` is the canonical command name. The colliding template is filtered
 		// out so autocomplete follows the interactive slash-command resolution path.
 		expect(matches.filter(name => name === "models")).toHaveLength(1);
 	});

packages/coding-agent/test/model-discovery.test.ts

@@ -1171,7 +1171,7 @@ describe("ModelRegistry runtime discovery", () => {
 	});
 
 	test("llama.cpp router preset refresh honors --ctx-size when the child hasn't been loaded yet", async () => {
-		// Reporter's workflow: `/model` picks a preset. On its very first switch
+		// Reporter's workflow: `/models` picks a preset. On its very first switch
 		// the child hasn't been spawned yet (meta.n_ctx absent), but the
 		// configured window is still what the user wants surfaced.
 		writeModelCache(

packages/coding-agent/test/model-registry.test.ts

@@ -451,7 +451,7 @@ describe("ModelRegistry", () => {
 
 		test("refresh keeps transport override on built-in provider (#2555 openrouter gateway)", async () => {
 			// Reporter ran `omp` with the auth-gateway broker proxying OpenRouter.
-			// Default model worked; switching via `/model` produced
+			// Default model worked; switching via `/models` produced
 			// `404 No route: POST /chat/completions` until restart. Root cause:
 			// background discovery refresh re-fetched the openrouter catalog and
 			// `mergeDiscoveredModel` dropped `transport: pi-native` (raw catalog

packages/coding-agent/test/modes/internal-url-autocomplete.test.ts

@@ -208,7 +208,7 @@ describe("internal-url-autocomplete", () => {
 
 		it("rejects non-url prefixes", () => {
 			expect(isInternalUrlPrefix("@src/foo")).toBe(false);
-			expect(isInternalUrlPrefix("/model")).toBe(false);
+			expect(isInternalUrlPrefix("/models")).toBe(false);
 		});
 	});
 

packages/coding-agent/test/prompt-action-autocomplete.test.ts

@@ -187,7 +187,7 @@ describe("prompt action autocomplete", () => {
 
 	it("delegates trySyncSlashCompletion to CombinedAutocompleteProvider", () => {
 		const provider = createPromptActionAutocompleteProvider({
-			commands: [{ name: "model", description: "Switch AI model" }],
+			commands: [{ name: "models", description: "Switch AI model" }],
 			basePath: "/tmp",
 			keybindings: AppKeybindingsManager.inMemory(),
 			copyCurrentLine: () => {},
@@ -201,12 +201,12 @@ describe("prompt action autocomplete", () => {
 
 		const result = provider.trySyncSlashCompletion("/mo");
 		expect(result).not.toBeNull();
-		expect(result!.items.map(i => i.value)).toContain("model");
+		expect(result!.items.map(i => i.value)).toContain("models");
 	});
 
 	it("returns null from trySyncSlashCompletion for non-slash text", () => {
 		const provider = createPromptActionAutocompleteProvider({
-			commands: [{ name: "model", description: "Switch AI model" }],
+			commands: [{ name: "models", description: "Switch AI model" }],
 			basePath: "/tmp",
 			keybindings: AppKeybindingsManager.inMemory(),
 			copyCurrentLine: () => {},

packages/coding-agent/test/slash-commands/history-security.test.ts

@@ -51,7 +51,7 @@ describe("shouldSkipHistory — security filter for slash command history", () =
 		expect(shouldSkipHistory("/plan do something")).toBe(false);
 		expect(shouldSkipHistory("/settings")).toBe(false);
 		expect(shouldSkipHistory("/btw what is this")).toBe(false);
-		expect(shouldSkipHistory("/model claude")).toBe(false);
+		expect(shouldSkipHistory("/models claude")).toBe(false);
 	});
 
 	it("returns false for non-slash text", () => {

packages/coding-agent/test/slash-commands/switch.test.ts

@@ -17,11 +17,11 @@ function createRuntime() {
 	};
 }
 
-describe("/model slash command", () => {
+describe("/models slash command", () => {
 	it("opens the model setup picker for role and thinking assignment", async () => {
 		const harness = createRuntime();
 
-		const handled = await executeBuiltinSlashCommand("/model", harness.runtime);
+		const handled = await executeBuiltinSlashCommand("/models", harness.runtime);
 
 		expect(handled).toBe(true);
 		expect(harness.showModelSelector.mock.calls).toEqual([[]]);