@mutagent/cli 0.1.150 → 0.1.152

package/dist/bin/cli.js

@@ -839,9 +839,10 @@ class SDKClientWrapper {
             maxIterations: config?.maxIterations ?? 1,
             targetScore: config?.targetScore ?? 0.8,
             patience: config?.patience,
-            model: config?.model,
-            ...config?.optimizerModel ? { optimizationModel: config.optimizerModel } : {},
+            ...config?.execModel ? { executionModel: config.execModel } : {},
+            ...config?.model ? { model: config.model } : {},
             ...config?.evalModel ? { evaluationModel: config.evalModel } : {},
+            ...config?.optModel ? { optimizationModel: config.optModel } : {},
             ...config?.providerId ? { executionProviderId: config.providerId } : {},
             ...config?.evalProviderId ? { evaluationProviderId: config.evalProviderId } : {},
             ...config?.optProviderId ? { optimizationProviderId: config.optProviderId } : {}
@@ -1107,6 +1108,17 @@ class SDKClientWrapper {
       this.handleError(error);
     }
   }
+  async getModelsCatalog() {
+    try {
+      const wsId = this.workspaceId;
+      if (!wsId)
+        return { models: [] };
+      const qs = `?workspaceId=${encodeURIComponent(wsId)}`;
+      return await this.request(`/api/providers/catalog${qs}`);
+    } catch {
+      return { models: [] };
+    }
+  }
   getCurrentWorkspaceId() {
     return this.workspaceId;
   }
@@ -6010,6 +6022,40 @@ ${chalk15.dim('Jobs with status "completed" or "cancelled" cannot be cancelled a
   });
 }
 
+// src/lib/optimizer-whitelist.ts
+var ALLOWED_OPTIMIZATION_MODELS = [
+  "gpt-5.2",
+  "gpt-5.4",
+  "claude-sonnet-4-6",
+  "claude-sonnet-4-6[1m]",
+  "claude-opus-4-6",
+  "claude-opus-4-6[1m]",
+  "anthropic.claude-sonnet-4-6",
+  "anthropic.claude-opus-4-6",
+  "anthropic.claude-opus-4-6-v1",
+  "anthropic.claude-sonnet-4-5-20250929-v1:0",
+  "claude-opus-4-5",
+  "anthropic.claude-opus-4-5-20251101-v1:0",
+  "gemini-3.1-pro-preview",
+  "gemini-3-flash-preview"
+];
+var WHITELIST_BY_PROVIDER = {
+  openai: ALLOWED_OPTIMIZATION_MODELS.filter((m) => m.startsWith("gpt-")),
+  anthropic: ALLOWED_OPTIMIZATION_MODELS.filter((m) => m.startsWith("claude-")),
+  bedrock: ALLOWED_OPTIMIZATION_MODELS.filter((m) => m.startsWith("anthropic.")),
+  google: ALLOWED_OPTIMIZATION_MODELS.filter((m) => m.startsWith("gemini-"))
+};
+function validateOptimizerModel(model) {
+  if (ALLOWED_OPTIMIZATION_MODELS.includes(model))
+    return;
+  const grouped = Object.entries(WHITELIST_BY_PROVIDER).filter(([, models]) => models.length > 0).map(([provider, models]) => `  ${provider}: ${models.join(", ")}`).join(`
+`);
+  throw new Error(`Model "${model}" is not allowed for the optimization slot.
+` + `Allowed optimization models by provider:
+${grouped}
+` + `Note: --exec-model and --eval-model are unrestricted.`);
+}
+
 // src/commands/prompts/optimize.ts
 var MODEL_FAMILIES = [
   {
@@ -6030,7 +6076,7 @@ var MODEL_FAMILIES = [
   {
     providerTypes: ["google"],
     prefixes: ["gemini-"],
-    examples: ["gemini-3.1-pro-preview", "gemini-3-flash-preview", "gemini-3.1-flash-lite-preview"]
+    examples: ["gemini-2.5-pro", "gemini-3.1-pro-preview", "gemini-3-flash-preview", "gemini-3.1-flash-lite-preview"]
   },
   {
     providerTypes: ["deepseek"],
@@ -6060,7 +6106,7 @@ ${chalk16.bold("Hit a bug or unexpected result?")}
     optimize.help();
   });
   prompts.addCommand(optimize);
-  optimize.command("start").description("Start prompt optimization").argument("<prompt-id>", "Prompt ID (from: mutagent prompts list)").requiredOption("-d, --dataset <id>", "Dataset ID for optimization (from: mutagent prompts dataset list <prompt-id>)").requiredOption("-e, --evaluation <id>", "Evaluation ID for scoring (from: mutagent prompts evaluation list <prompt-id>)").option("--max-iterations <n>", "Max optimization iterations (default: 1)").option("--target-score <n>", "Target accuracy 0-1 (default: 0.8)").option("--patience <n>", "Iterations without improvement before stopping").option("--model <model-id>", 'Target LLM model (e.g., "claude-sonnet-4-5-20250929")').option("--eval-model <model-id>", "Evaluation model (defaults to target model)").option("--optimizer-model <model-id>", "Optimization/mutation model (defaults to eval model)").option("--provider-id <id>", "Provider config ID for the execution slot (provider-id-first routing)").option("--eval-provider-id <id>", "Provider config ID for the evaluation slot").option("--opt-provider-id <id>", "Provider config ID for the optimization slot").option("--watch", "Watch live progress with stage cards", false).addHelpText("after", `
+  optimize.command("start").description("Start prompt optimization").argument("<prompt-id>", "Prompt ID (from: mutagent prompts list)").requiredOption("-d, --dataset <id>", "Dataset ID for optimization (from: mutagent prompts dataset list <prompt-id>)").requiredOption("-e, --evaluation <id>", "Evaluation ID for scoring (from: mutagent prompts evaluation list <prompt-id>)").option("--max-iterations <n>", "Max optimization iterations (default: 1)").option("--target-score <n>", "Target accuracy 0-1 (default: 0.8)").option("--patience <n>", "Iterations without improvement before stopping").option("--model <model-id>", 'Execution model (e.g., "claude-sonnet-4-5-20250929")').option("--exec-model <model-id>", "Execution model — alias for --model; takes precedence if both provided").option("--eval-model <model-id>", "Evaluation model — unrestricted (defaults to exec model)").option("--opt-model <model-id>", "Optimizer model — must be on the optimizer whitelist").option("--optimizer-model <model-id>", "[deprecated] Alias for --opt-model. Use --opt-model instead.").option("--provider-id <id>", "Provider config ID for the execution slot (provider-id-first routing)").option("--eval-provider-id <id>", "Provider config ID for the evaluation slot").option("--opt-provider-id <id>", "Provider config ID for the optimization slot").option("--watch", "Watch live progress with stage cards", false).addHelpText("after", `
 Examples:
   ${chalk16.dim("$")} mutagent prompts optimize start <prompt-id> --dataset <dataset-id> --evaluation <eval-id>
   ${chalk16.dim("$")} mutagent prompts optimize start <prompt-id> --dataset <dataset-id> --evaluation <eval-id> --max-iterations 5
@@ -6096,10 +6142,19 @@ ${chalk16.bold("Hit a bug or unexpected result?")}
     try {
       await checkProviderConfigured();
       const client = await getSDKClient();
-      if (!options.model && !isJson) {
+      if (options.optimizerModel) {
+        if (!isJson)
+          output.warn("--optimizer-model is deprecated, use --opt-model");
+        options.optModel = options.optModel ?? options.optimizerModel;
+      }
+      if (options.optModel) {
+        validateOptimizerModel(options.optModel);
+      }
+      const execModel = options.execModel ?? options.model;
+      if (!execModel && !isJson) {
         output.warn("No --model specified. Server will use default. Pass --model <id> to control costs.");
       }
-      if (options.model) {
+      if (execModel) {
         try {
           const providers = await client.listProviders();
           if (providers.data && providers.data.length > 0) {
@@ -6112,11 +6167,11 @@ ${chalk16.bold("Hit a bug or unexpected result?")}
               }
             } else {
               const modelFamilies = getModelFamiliesForProviderTypes(configuredTypes);
-              const modelLower = options.model.toLowerCase();
+              const modelLower = execModel.toLowerCase();
               const matchedFamily = modelFamilies.find(({ prefixes }) => prefixes.some((prefix) => modelLower.startsWith(prefix)));
               if (!matchedFamily) {
                 const supportedModels = modelFamilies.flatMap((f) => f.examples).join(", ");
-                const errorMsg = `Model '${options.model}' is not supported by any configured provider. ` + `Your providers: [${configuredTypes.join(", ")}]. ` + `Supported models: ${supportedModels || "(unknown — check mutagent providers list --json)"}. ` + `Run: mutagent providers list --json`;
+                const errorMsg = `Model '${execModel}' is not supported by any configured provider. ` + `Your providers: [${configuredTypes.join(", ")}]. ` + `Supported models: ${supportedModels || "(unknown — check mutagent providers list --json)"}. ` + `Run: mutagent providers list --json`;
                 if (isJson) {
                   output.output({ success: false, error: errorMsg, code: "MODEL_NOT_SUPPORTED" });
                   return;
@@ -6280,9 +6335,10 @@ ${chalk16.bold("Hit a bug or unexpected result?")}
         maxIterations: options.maxIterations ? parseInt(options.maxIterations, 10) : 1,
         targetScore: options.targetScore ? parseFloat(options.targetScore) : undefined,
         patience: options.patience ? parseInt(options.patience, 10) : undefined,
-        model: options.model,
+        model: options.execModel ? undefined : options.model,
+        execModel: options.execModel,
         evalModel: options.evalModel,
-        optimizerModel: options.optimizerModel,
+        optModel: options.optModel,
         providerId: options.providerId,
         evalProviderId: options.evalProviderId,
         optProviderId: options.optProviderId
@@ -8810,11 +8866,15 @@ Provider Types:
 
 Subcommands:
   list, get, add, update, delete, test
+
+Tip:
+  Use --models on list to see available models per provider.
     `);
-  providers.command("list").description("List all providers").option("-l, --limit <n>", "Limit results", "50").option("-o, --offset <n>", "Offset for pagination").option("-t, --type <type>", "Filter by provider type").addHelpText("after", `
+  providers.command("list").description("List all providers").option("-l, --limit <n>", "Limit results", "50").option("-o, --offset <n>", "Offset for pagination").option("-t, --type <type>", "Filter by provider type").option("-m, --models", "Show available models per provider (calls /providers/catalog)").addHelpText("after", `
 Examples:
   ${chalk33.dim("$")} mutagent providers list
   ${chalk33.dim("$")} mutagent providers list --type openai
+  ${chalk33.dim("$")} mutagent providers list --models
   ${chalk33.dim("$")} mutagent providers list --json
     `).action(async (options) => {
     const isJson = getJsonFlag(providers);
@@ -8831,6 +8891,26 @@ Examples:
         filters.type = validateProviderType(options.type);
       }
       const result = await client.listProviders(filters);
+      let catalogByKind = {};
+      if (options.models) {
+        try {
+          const catalog = await client.getModelsCatalog();
+          for (const entry of catalog.models) {
+            const kind = entry.provider.kind;
+            if (!catalogByKind[kind])
+              catalogByKind[kind] = [];
+            catalogByKind[kind].push(entry.id);
+          }
+        } catch {}
+      }
+      const formatModels = (kind) => {
+        const models = catalogByKind[kind] ?? [];
+        if (models.length === 0)
+          return "—";
+        if (models.length <= 3)
+          return models.join(", ");
+        return `${models.slice(0, 3).join(", ")} +${models.length - 3} more`;
+      };
       if (isJson) {
         const withLinks = result.data.map((p) => ({
           id: p.id,
@@ -8838,6 +8918,7 @@ Examples:
           type: p.type,
           isActive: p.isActive,
           updatedAt: p.updatedAt,
+          ...options.models ? { models: catalogByKind[p.type] ?? [] } : {},
           _links: providerLinks(p.id)
         }));
         output.output({ ...result, data: withLinks });
@@ -8853,6 +8934,7 @@ Examples:
             baseUrl: p.baseUrl ?? "default",
             active: p.isActive ? "Yes" : "No",
             updated: p.updatedAt ? new Date(p.updatedAt).toLocaleDateString() : "N/A",
+            ...options.models ? { models: formatModels(p.type) } : {},
             url: providerLink(p.id)
           }));
           output.output(formatted);
@@ -9544,6 +9626,7 @@ description: |
 3. **NEVER auto-generate eval criteria — collect from user.** Ask the user for each rubric field. See [concepts/eval-criteria.md](./concepts/eval-criteria.md) for the 3-tier format.
 4. **Explore-before-modify.** Run \`mutagent explore --json\` before any write operation. Present findings, get user confirmation. Never mutate without discovery first.
 5. **Cost transparency before \`optimize start\`.** Run \`mutagent usage --json\` and show the result to the user. Get explicit confirmation before any optimization job.
+6. **Before optimizing, run \`mutagent providers list --models\` to verify available models.** This calls \`/providers/catalog\` and shows which models are available per provider. Use the output to pick valid \`--exec-model\` and \`--eval-model\` values.
 
 ---
 
@@ -9559,6 +9642,7 @@ Match the user's first request. Load ONLY the matching subfile. Do NOT preload t
 | \`AgentExecutor\`, \`StateGraph\`, \`createReactAgent\`, \`tool_calls\`, \`@tool\`, \`langgraph\`, \`crewai\`, \`autogen\`, \`openai/agents\`, multi-turn | [workflows/agents.md](./workflows/agents.md) | WIP path — surface partnership link |
 | "how do variables work", "single vs double braces", delimiter | [concepts/prompt-variables.md](./concepts/prompt-variables.md) | Delimiter inference contract |
 | "what makes a good eval", "rubric", "evaluation criteria" | [concepts/eval-criteria.md](./concepts/eval-criteria.md) | INPUT MVC + OUTPUT Standards |
+| "check models", "what models", "available models", "which models" | run \`mutagent providers list --models --json\` | Discovery: shows catalog per provider before model selection |
 | Unclear / first time | run \`mutagent explore --json\` first, then reroute | Discovery before action |
 
 ---
@@ -12150,6 +12234,9 @@ ${chalk39.yellow("Command Navigation:")}
   mutagent prompts evaluation create <id> --guided --json   ${chalk39.dim("Guided eval workflow")}
   mutagent prompts evaluation list <id> --json        ${chalk39.dim("List evaluations")}
 
+  mutagent providers list --json              ${chalk39.dim("List configured BYOK providers")}
+  mutagent providers list --models            ${chalk39.dim("see available models per provider")}
+
   mutagent prompts optimize start --help      ${chalk39.dim("Run optimization (read help first!)")}
   mutagent prompts optimize status <job-id>   ${chalk39.dim("Poll progress")}
   mutagent prompts optimize results <job-id>  ${chalk39.dim("View scorecard")}
@@ -12292,5 +12379,5 @@ if (isInteractive && !isSkillCommand) {
 }
 program.parse();
 
-//# debugId=2FD2902EDF36086764756E2164756E21
+//# debugId=EE09049CBBE922DA64756E2164756E21
 //# sourceMappingURL=cli.js.map