fix: bound AI analysis provider latency

modules/ai/server/config.ts

@@ -3,6 +3,9 @@ export type AiRuntimeConfig = {
   baseUrl: string;
   apiKey: string;
   chatModel: string;
+  requestTimeoutMs: number;
+  maxTokens: number;
+  maxRetries: number;
 };
 
 export type AiRuntimeConfigResult =
@@ -10,11 +13,28 @@ export type AiRuntimeConfigResult =
   | { success: false; reason: string };
 
 const DEFAULT_CHAT_MODEL = "gpt-4.1-mini";
+const DEFAULT_REQUEST_TIMEOUT_MS = 20_000;
+const DEFAULT_MAX_TOKENS = 1_000;
+const DEFAULT_MAX_RETRIES = 0;
 
 function readOptionalEnv(name: string): string {
   return process.env[name]?.trim() ?? "";
 }
 
+function readBoundedIntegerEnv(name: string, fallback: number, min: number, max: number): number {
+  const rawValue = readOptionalEnv(name);
+  if (!rawValue) {
+    return fallback;
+  }
+
+  const value = Number(rawValue);
+  if (!Number.isInteger(value) || value < min || value > max) {
+    return fallback;
+  }
+
+  return value;
+}
+
 export function getAiRuntimeConfig(): AiRuntimeConfigResult {
   const apiKey = readOptionalEnv("AI_API_KEY");
   if (!apiKey) {
@@ -27,6 +47,10 @@ export function getAiRuntimeConfig(): AiRuntimeConfigResult {
       baseUrl: readOptionalEnv("AI_BASE_URL"),
       apiKey,
       chatModel: readOptionalEnv("AI_CHAT_MODEL") || DEFAULT_CHAT_MODEL,
+      requestTimeoutMs: readBoundedIntegerEnv("AI_REQUEST_TIMEOUT_MS", DEFAULT_REQUEST_TIMEOUT_MS, 1_000, 120_000),
+      maxTokens: readBoundedIntegerEnv("AI_MAX_TOKENS", DEFAULT_MAX_TOKENS, 256, 2_000),
+      maxRetries: readBoundedIntegerEnv("AI_MAX_RETRIES", DEFAULT_MAX_RETRIES, 0, 5),
     },
   };
 }
+