wolverine-ai 3.7.3 → 3.7.5

package/bin/wolverine.js

@@ -144,9 +144,10 @@ console.log(chalk.yellow.bold("\n  🐺 Wolverine Node.js — Autonomous Server
 logSystemInfo(systemInfo);
 console.log("");
 
-console.log(chalk.bold("  Models:"));
+console.log(chalk.bold(`  Models (${config.provider}):`));
 logModelConfig(chalk);
 console.log("");
+console.log(chalk.gray(`  Provider:   ${config.provider}`));
 console.log(chalk.gray(`  Script:     ${scriptPath}`));
 console.log(chalk.gray(`  Port:       ${config.server.port}`));
 console.log(chalk.gray(`  Retries:    ${config.server.maxRetries}`));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "wolverine-ai",
-  "version": "3.7.3",
+  "version": "3.7.5",
   "description": "Self-healing Node.js server framework powered by AI. Catches crashes, diagnoses errors, generates fixes, verifies, and restarts — automatically.",
   "main": "src/index.js",
   "bin": {

package/src/agent/agent-engine.js

@@ -414,7 +414,7 @@ class AgentEngine {
       }
 
       let response;
-      const AI_CALL_TIMEOUT_MS = 45000; // 45s per API call — prevents indefinite hangs
+      const AI_CALL_TIMEOUT_MS = parseInt(process.env.WOLVERINE_AI_CALL_TIMEOUT_MS, 10) || 90000; // 90s per API call — self-hosted GPU needs more time
       try {
         response = await Promise.race([
           aiCallWithHistory({
@@ -423,7 +423,7 @@ class AgentEngine {
             tools: allTools,
             maxTokens: 4096,
           }),
-          new Promise((_, reject) => setTimeout(() => reject(new Error("AI call timed out after 45s")), AI_CALL_TIMEOUT_MS)),
+          new Promise((_, reject) => setTimeout(() => reject(new Error(`AI call timed out after ${AI_CALL_TIMEOUT_MS / 1000}s`)), AI_CALL_TIMEOUT_MS)),
         ]);
       } catch (err) {
         console.log(chalk.red(`  Agent API error: ${err.message}`));

package/src/core/ai-client.js

@@ -169,7 +169,20 @@ async function _withRetry(fn, maxRetries = 3) {
     try {
       return await fn();
     } catch (err) {
-      const isRateLimit = err.status === 429 || err.code === "rate_limit_exceeded";
+      const msg = (err.message || "").toLowerCase();
+      const code = (err.code || "").toLowerCase();
+
+      // Permanent billing/quota errors — never retry, surface immediately
+      const isBillingError = err.status === 402
+        || /insufficient.*(quota|credits|funds)/i.test(msg)
+        || /billing_hard_limit|insufficient_quota|quota_exceeded/i.test(msg)
+        || /billing_hard_limit|insufficient_quota|quota_exceeded/i.test(code);
+      if (isBillingError) {
+        console.log(chalk.red(`  💳 Billing error (not retrying): ${err.message}`));
+        throw err;
+      }
+
+      const isRateLimit = err.status === 429 || code === "rate_limit_exceeded";
       const isServerError = err.status >= 500;
       if ((isRateLimit || isServerError) && attempt < maxRetries) {
         const delay = Math.min(1000 * Math.pow(2, attempt) + Math.random() * 1000, 30000);

package/src/core/config.js

@@ -43,6 +43,7 @@ function loadConfig() {
       classifier: process.env.CLASSIFIER_MODEL   || modelSource.classifier || "gpt-4o-mini",
       audit:      process.env.AUDIT_MODEL        || modelSource.audit      || "gpt-4o-mini",
       compacting: process.env.COMPACTING_MODEL   || modelSource.compacting || "gpt-4o-mini",
+      utility:    process.env.COMPACTING_MODEL   || modelSource.compacting || "gpt-4o-mini",
       research:   process.env.RESEARCH_MODEL     || modelSource.research   || "gpt-4o",
       embedding:  process.env.TEXT_EMBEDDING_MODEL || modelSource.embedding || "text-embedding-3-small",
     },

package/src/core/models.js

@@ -119,9 +119,12 @@ function getModel(role) {
 function getModelConfig() {
   const config = {};
   for (const [role, def] of Object.entries(MODEL_ROLES)) {
+    const resolved = getModel(role);
+    const fromEnv = !!process.env[def.envKey];
+    const fromDefault = resolved === def.default;
     config[role] = {
-      model: process.env[def.envKey] || def.default,
-      source: process.env[def.envKey] ? "env" : "default",
+      model: resolved,
+      source: fromEnv ? "env" : fromDefault ? "default" : "settings",
       tier: def.tier,
     };
   }

package/src/notifications/notifier.js

@@ -29,11 +29,17 @@ const HUMAN_REQUIRED_PATTERNS = [
   { pattern: /(api|auth|token|key|credential).*(expired|revoked|rotated|invalid)/i, category: "auth", hint: "Credential has expired or been revoked" },
   { pattern: /authentication\s+failed/i, category: "auth", hint: "Authentication failed — check credentials" },
 
-  // Billing/Quota
+  // Billing/Quota — covers OpenAI, Anthropic, Wolverine, and generic patterns
   { pattern: /429\s*(too many|rate limit)/i, category: "billing", hint: "Rate limit hit — may need to upgrade plan or wait" },
   { pattern: /(quota|limit|credits?)\s*(exceeded|exhausted|depleted)/i, category: "billing", hint: "Usage quota or credits exhausted" },
   { pattern: /billing.*(?:issue|error|failed|inactive)/i, category: "billing", hint: "Billing issue on the account" },
   { pattern: /insufficient.*(funds|credits|quota)/i, category: "billing", hint: "Insufficient credits or funds" },
+  { pattern: /billing_hard_limit_reached/i, category: "billing", hint: "OpenAI billing hard limit reached — add payment method or raise limit" },
+  { pattern: /insufficient_quota/i, category: "billing", hint: "API quota exhausted — check billing dashboard" },
+  { pattern: /rate_limit_exceeded/i, category: "billing", hint: "API rate limit exceeded — wait or upgrade plan" },
+  { pattern: /402\s*(payment|required)/i, category: "billing", hint: "Payment required — check billing status" },
+  { pattern: /exceeded.*(?:budget|spending|token)/i, category: "billing", hint: "Spending or token budget exceeded" },
+  { pattern: /overloaded_error/i, category: "billing", hint: "Anthropic API overloaded — retry later" },
 
   // External service failures
   { pattern: /ECONNREFUSED/i, category: "service", hint: "External service connection refused — is it running?" },