@vellumai/assistant 0.5.1 → 0.5.2

package/src/daemon/conversation-slash.ts

@@ -4,18 +4,14 @@ import { join } from "node:path";
 
 import QRCode from "qrcode";
 
+import type { InterfaceId } from "../channels/types.js";
 import { getGatewayPort, getIngressPublicBaseUrl } from "../config/env.js";
-import { getConfig, loadRawConfig, saveRawConfig } from "../config/loader.js";
-import { setServiceField } from "../config/raw-config-utils.js";
-import {
-  getConfiguredProviders,
-  isProviderAvailable,
-} from "../providers/provider-availability.js";
-import { initializeProviders } from "../providers/registry.js";
+import { getConfig } from "../config/loader.js";
+import { PROVIDER_CATALOG } from "../providers/model-catalog.js";
+import { getConfiguredProviders } from "../providers/provider-availability.js";
 import { getLocalIPv4 } from "../util/network-info.js";
 import { getWorkspaceDir } from "../util/platform.js";
 import { silentlyWithLog } from "../util/silently.js";
-import { getAssistantName } from "./identity-helpers.js";
 import type { PairingStore } from "./pairing-store.js";
 
 export type SlashResolution =
@@ -45,160 +41,51 @@ export interface SlashContext {
   model: string;
   provider: string;
   estimatedCost: number;
+  userMessageInterface?: InterfaceId;
 }
 
-// ── /model command ───────────────────────────────────────────────────
-
-const AVAILABLE_MODELS = [
-  "claude-opus-4-6",
-  "claude-sonnet-4-6",
-  "claude-haiku-4-5-20251001",
-] as const;
-
-const MODEL_DISPLAY_NAMES: Record<string, string> = {
-  "claude-opus-4-6": "Claude Opus 4.6",
-  "claude-sonnet-4-6": "Claude Sonnet 4.6",
-  "claude-haiku-4-5-20251001": "Claude Haiku 4.5",
-};
-
-const PROVIDER_MODEL_SHORTCUTS: Record<
-  string,
-  { provider: string; model: string; displayName: string }
-> = {
-  // Anthropic
-  opus: {
-    provider: "anthropic",
-    model: "claude-opus-4-6",
-    displayName: "Claude Opus 4.6",
-  },
-  sonnet: {
-    provider: "anthropic",
-    model: "claude-sonnet-4-6",
-    displayName: "Claude Sonnet 4.6",
-  },
-  haiku: {
-    provider: "anthropic",
-    model: "claude-haiku-4-5-20251001",
-    displayName: "Claude Haiku 4.5",
-  },
-  "grok-beta": {
-    provider: "openrouter",
-    model: "x-ai/grok-4.20-beta",
-    displayName: "Grok 4.20 Beta (OpenRouter)",
-  },
-  "grok-multi": {
-    provider: "openrouter",
-    model: "x-ai/grok-4.20-beta",
-    displayName: "Grok 4.20 Beta (OpenRouter)",
-  },
-};
-
-/** True when the trimmed content matches a provider shortcut like /opus, /gpt4, etc. */
-export function isProviderShortcut(content: string): boolean {
-  const match = content.trim().match(/^\/([a-z0-9-]+)(\s|$)/i);
-  if (!match) return false;
-  return match[1].toLowerCase() in PROVIDER_MODEL_SHORTCUTS;
-}
-
-/** Reverse lookup: model ID → provider, derived from PROVIDER_MODEL_SHORTCUTS. */
-export const MODEL_TO_PROVIDER: Record<string, string> = Object.fromEntries(
-  Object.values(PROVIDER_MODEL_SHORTCUTS).map(({ model, provider }) => [
-    model,
-    provider,
-  ]),
-);
-
-/** Partial-match a user input like "opus", "sonnet", "haiku" to a full model ID. */
-function matchModel(input: string): string | undefined {
-  const lower = input.toLowerCase().trim();
-  // Exact match first
-  const exact = AVAILABLE_MODELS.find((m) => m === lower);
-  if (exact) return exact;
-  // Partial match (e.g. "opus" → "claude-opus-4-6")
-  return AVAILABLE_MODELS.find((m) => m.includes(lower));
-}
-
-async function resolveProviderModelCommand(
-  content: string,
-): Promise<SlashResolution | null> {
-  const trimmed = content.trim();
-  if (!trimmed.startsWith("/")) return null;
+// ── Deprecated model-switching shortcuts ─────────────────────────────
 
-  // Extract the command (e.g., "/gpt4" → "gpt4")
-  const match = trimmed.match(/^\/([a-z0-9-]+)(\s|$)/i);
-  if (!match) return null;
-
-  const command = match[1].toLowerCase();
-  const shortcut = PROVIDER_MODEL_SHORTCUTS[command];
-  if (!shortcut) return null;
-
-  const { provider, model, displayName } = shortcut;
-  const config = getConfig();
-  const name = getAssistantName();
-
-  // Check if provider is available (secure key, env var, managed proxy, or no key needed)
-  if (!(await isProviderAvailable(provider))) {
-    return {
-      kind: "unknown",
-      message: `Cannot switch to ${displayName}. No API key configured for ${provider}.\n\nSet it with: \`keys set ${provider} <your-key>\``,
-    };
-  }
-
-  // Check if already using this provider+model
-  if (
-    config.services.inference.provider === provider &&
-    config.services.inference.model === model
-  ) {
-    const alreadyMsg = name
-      ? `${name} is already running on **${displayName}**.`
-      : `Already using **${displayName}**.`;
-    return {
-      kind: "unknown",
-      message: alreadyMsg,
-    };
-  }
-
-  // Update config with both provider and model
-  const raw = loadRawConfig();
-  setServiceField(raw, "inference", "provider", provider);
-  setServiceField(raw, "inference", "model", model);
-  saveRawConfig(raw);
-
-  // Re-initialize providers with new config
-  const newConfig = getConfig();
-  await initializeProviders(newConfig);
+/**
+ * Former provider shortcut commands that switched models. These are now
+ * removed — model switching lives in Settings. We reject them explicitly
+ * so they don't fall through to the LLM as passthrough text.
+ */
+const DEPRECATED_MODEL_SHORTCUTS = new Set([
+  "opus",
+  "sonnet",
+  "haiku",
+  "grok-beta",
+  "grok-multi",
+]);
 
-  const switchedMsg = name
-    ? `Switched ${name} to **${displayName}**. New conversations will use this model.`
-    : `Switched to **${displayName}**. New conversations will use this model.`;
-
-  return {
-    kind: "unknown",
-    message: switchedMsg,
-  };
-}
+// ── /models command ──────────────────────────────────────────────────
 
 async function resolveModelList(): Promise<SlashResolution> {
   const config = getConfig();
-
-  // Build a set of providers that are usable (secure key, env var, or managed proxy).
   const configuredProviders = new Set<string>(await getConfiguredProviders());
 
   const lines = ["Available models:\n"];
 
-  for (const [cmd, { provider, model, displayName }] of Object.entries(
-    PROVIDER_MODEL_SHORTCUTS,
-  )) {
+  for (const {
+    id: provider,
+    displayName: providerName,
+    models,
+  } of PROVIDER_CATALOG) {
     const hasKey = configuredProviders.has(provider);
-    const isCurrent =
-      config.services.inference.provider === provider &&
-      config.services.inference.model === model;
     const status = hasKey ? "✓" : "✗";
-    const current = isCurrent ? " **[current]**" : "";
-    lines.push(`- **${displayName}** (/${cmd}) ${status}${current}`);
+    lines.push(`**${providerName}** ${status}`);
+    for (const { id, displayName } of models) {
+      const isCurrent =
+        config.services.inference.provider === provider &&
+        config.services.inference.model === id;
+      const current = isCurrent ? " **[current]**" : "";
+      lines.push(`  - ${displayName} (\`${id}\`)${current}`);
+    }
+    lines.push("");
   }
 
-  lines.push("\n✓ = API key configured, ✗ = not configured");
+  lines.push("✓ = API key configured, ✗ = not configured");
   lines.push("\nTip: Configure a provider with `keys set <provider> <key>`");
 
   return {
@@ -207,92 +94,6 @@ async function resolveModelList(): Promise<SlashResolution> {
   };
 }
 
-async function resolveModelCommand(
-  content: string,
-): Promise<SlashResolution | null> {
-  const trimmed = content.trim();
-  // Match /models → route to list
-  if (trimmed === "/models") {
-    return await resolveModelList();
-  }
-
-  if (!trimmed.startsWith("/model")) return null;
-  // Ensure it's exactly "/model" or "/model " (not "/modelsomething")
-  if (trimmed.length > 6 && trimmed[6] !== " ") return null;
-
-  const args = trimmed.slice(6).trim();
-  const name = getAssistantName();
-
-  if (!args) {
-    // Show current model
-    const config = getConfig();
-    const displayName =
-      MODEL_DISPLAY_NAMES[config.services.inference.model] ??
-      config.services.inference.model;
-    const prefix = name ? `${name} is running on` : `Currently using`;
-    return {
-      kind: "unknown",
-      message: `${prefix} **${displayName}** (\`${config.services.inference.model}\`).`,
-    };
-  }
-
-  // Handle /model list
-  if (args === "list") {
-    return await resolveModelList();
-  }
-
-  // Try to match the model name
-  const matched = matchModel(args);
-  if (!matched) {
-    const available = AVAILABLE_MODELS.map(
-      (m) => `- **${MODEL_DISPLAY_NAMES[m]}** (\`${m}\`)`,
-    ).join("\n");
-    return {
-      kind: "unknown",
-      message: `Hmm, "${args}" doesn't match any available model. Here's what you can pick from:\n${available}`,
-    };
-  }
-
-  // Check if already using this model
-  const currentConfig = getConfig();
-  if (currentConfig.services.inference.model === matched) {
-    const displayName = MODEL_DISPLAY_NAMES[matched] ?? matched;
-    const alreadyMsg = name
-      ? `${name} is already running on **${displayName}**.`
-      : `Already on **${displayName}**.`;
-    return {
-      kind: "unknown",
-      message: alreadyMsg,
-    };
-  }
-
-  // Validate that Anthropic provider is available (secure key, env var, or managed proxy)
-  if (!(await isProviderAvailable("anthropic"))) {
-    const displayName = MODEL_DISPLAY_NAMES[matched] ?? matched;
-    return {
-      kind: "unknown",
-      message: `Cannot switch to ${displayName}. No API key configured for Anthropic.\n\nSet it with: \`keys set anthropic <your-key>\``,
-    };
-  }
-
-  // Change model: save config and re-initialize providers
-  const raw = loadRawConfig();
-  setServiceField(raw, "inference", "provider", "anthropic"); // Ensure provider is set for Anthropic models
-  setServiceField(raw, "inference", "model", matched);
-  saveRawConfig(raw);
-  const config = getConfig();
-  await initializeProviders(config);
-
-  const displayName = MODEL_DISPLAY_NAMES[matched] ?? matched;
-  const switchedMsg = name
-    ? `Switched ${name} to **${displayName}**. New conversations will use this model.`
-    : `Switched to **${displayName}**. New conversations will use this model.`;
-  return {
-    kind: "unknown",
-    message: switchedMsg,
-  };
-}
-
 function resolveStatusCommand(context: SlashContext): SlashResolution {
   const {
     inputTokens,
@@ -310,14 +111,13 @@ function resolveStatusCommand(context: SlashContext): SlashResolution {
   const filled = Math.round(pct / 5);
   const bar = "█".repeat(filled) + "░".repeat(20 - filled);
   const fmt = (n: number) => n.toLocaleString("en-US");
-  const displayName = MODEL_DISPLAY_NAMES[model] ?? model;
 
   const lines = [
     "Conversation Status\n",
     `Context:  ${bar}  ${pct}%  (${fmt(inputTokens)} / ${fmt(
       maxInputTokens,
     )} tokens)`,
-    `Model:    ${displayName} (${provider})`,
+    `Model:    ${model} (${provider})`,
     `Messages: ${fmt(messageCount)}`,
     `Tokens:   ${fmt(inputTokens)} in / ${fmt(outputTokens)} out`,
     `Cost:     $${estimatedCost.toFixed(2)} (estimated)`,
@@ -326,28 +126,91 @@ function resolveStatusCommand(context: SlashContext): SlashResolution {
   return { kind: "unknown", message: lines.join("\n") };
 }
 
+function resolveCommandsList(context?: SlashContext): string[] {
+  const fallbackLines = [
+    "/commands — List all available commands",
+    "/models — List all available models",
+    "/pair — Generate pairing info for connecting a mobile device",
+  ];
+  if (context) {
+    fallbackLines.push("/status — Show conversation status and context usage");
+  }
+
+  if (!context?.userMessageInterface) return fallbackLines;
+
+  if (context.userMessageInterface === "macos") {
+    return [
+      "/commands — List all available commands",
+      "/models — List all available models",
+      "/status — Show conversation status and context usage",
+      "/btw — Ask a side question while the assistant is working",
+      "/fork — Fork the current conversation into a new branch",
+      "/pair — Generate pairing info for connecting a mobile device",
+    ];
+  }
+
+  if (context.userMessageInterface === "ios") {
+    return [
+      "/commands — List all available commands",
+      "/models — List all available models",
+      "/status — Show conversation status and context usage",
+      "/btw — Ask a side question while the assistant is working",
+      "/fork — Fork the current conversation into a new branch",
+    ];
+  }
+
+  return [
+    "/commands — List all available commands",
+    "/models — List all available models",
+    "/status — Show conversation status and context usage",
+    "/btw — Ask a side question while the assistant is working",
+  ];
+}
+
 /**
- * Resolve built-in slash commands (/model, /status, /commands, /pair, provider shortcuts).
+ * Resolve built-in slash commands (/models, /status, /commands, /pair).
  * Returns `unknown` with a deterministic message, or the (possibly rewritten) content.
  */
 export async function resolveSlash(
   content: string,
   context?: SlashContext,
 ): Promise<SlashResolution> {
-  // Check provider shortcuts first (/gpt4, /opus, etc.)
-  const providerResult = await resolveProviderModelCommand(content);
-  if (providerResult) return providerResult;
+  // Handle deprecated model-switching commands — direct users to Settings
+  const trimmed = content.trim();
+  if (
+    trimmed === "/model" ||
+    (trimmed.startsWith("/model ") && trimmed !== "/models")
+  ) {
+    return {
+      kind: "unknown",
+      message:
+        "The `/model` command has been removed. Use **Settings → Models & Services** to change your model and provider.",
+    };
+  }
 
-  // Handle /model command
-  const modelResult = await resolveModelCommand(content);
-  if (modelResult) return modelResult;
+  // Reject deprecated provider shortcut commands (/opus, /sonnet, /haiku, etc.)
+  const shortcutMatch = trimmed.match(/^\/([a-z0-9-]+)(\s|$)/i);
+  if (
+    shortcutMatch &&
+    DEPRECATED_MODEL_SHORTCUTS.has(shortcutMatch[1].toLowerCase())
+  ) {
+    return {
+      kind: "unknown",
+      message: `The \`/${shortcutMatch[1]}\` shortcut has been removed. Use **Settings → Models & Services** to change your model and provider.`,
+    };
+  }
+
+  // Handle /models command (read-only listing)
+  if (trimmed === "/models") {
+    return await resolveModelList();
+  }
 
   // Handle /pair command
-  const pairResult = resolvePairCommand(content);
+  const pairResult = resolvePairCommand(content, context);
   if (pairResult) return pairResult;
 
   // Handle /status command
-  if (content.trim() === "/status") {
+  if (trimmed === "/status") {
     if (!context) {
       return {
         kind: "unknown",
@@ -358,19 +221,10 @@ export async function resolveSlash(
   }
 
   // Handle /commands command
-  if (content.trim() === "/commands") {
-    const lines = [
-      "/commands — List all available commands",
-      "/model — Show or switch the current model",
-      "/models — List all available models",
-      "/pair — Generate pairing info for connecting a mobile device",
-    ];
-    if (context) {
-      lines.push("/status — Show conversation status and context usage");
-    }
+  if (trimmed === "/commands") {
     return {
       kind: "unknown",
-      message: lines.join("\n"),
+      message: resolveCommandsList(context).join("\n"),
     };
   }
 
@@ -406,9 +260,20 @@ async function savePairingQRCodePng(
   writeFileSync(qrPngPath, pngBuffer);
 }
 
-function resolvePairCommand(content: string): SlashResolution | null {
+function resolvePairCommand(
+  content: string,
+  context?: SlashContext,
+): SlashResolution | null {
   if (content.trim() !== "/pair") return null;
 
+  if (context?.userMessageInterface && context.userMessageInterface !== "macos") {
+    return {
+      kind: "unknown",
+      message:
+        "The `/pair` command is only available in the macOS desktop app.",
+    };
+  }
+
   if (!pairingStoreRef) {
     return {
       kind: "unknown",