lcclaude 1.0.0 → 1.0.1

package/.env.example

@@ -1,18 +1,15 @@
-# Default example: Anthropic-compatible gateway (e.g. MiniMax / OpenRouter / proxy)
-ANTHROPIC_AUTH_TOKEN=your_token_here
-ANTHROPIC_BASE_URL=https://api.minimaxi.com/anthropic
-ANTHROPIC_DEFAULT_HAIKU_MODEL=MiniMax-M2.7-highspeed
-ANTHROPIC_DEFAULT_OPUS_MODEL=MiniMax-M2.7-highspeed
-ANTHROPIC_DEFAULT_SONNET_MODEL=MiniMax-M2.7-highspeed
-ANTHROPIC_MODEL=MiniMax-M2.7-highspeed
+# Default example: local Qwen / OpenAI-compatible backend via LiteLLM
+ANTHROPIC_AUTH_TOKEN=dummy
+ANTHROPIC_BASE_URL=http://127.0.0.1:4000/anthropic
+ANTHROPIC_MODEL=qwen-local
+ANTHROPIC_DEFAULT_SONNET_MODEL=qwen-local
+ANTHROPIC_DEFAULT_HAIKU_MODEL=qwen-local
+ANTHROPIC_DEFAULT_OPUS_MODEL=qwen-local
+CLAUDE_CODE_MAX_OUTPUT_TOKENS=4096
+CLAUDE_CODE_MAX_CONTEXT_TOKENS=65536
+CLAUDE_CODE_TEMPERATURE=0
 
-# Local Qwen / OpenAI-compatible example via LiteLLM:
-# ANTHROPIC_AUTH_TOKEN=dummy
-# ANTHROPIC_BASE_URL=http://127.0.0.1:4000/anthropic
-# ANTHROPIC_MODEL=qwen-local
-# ANTHROPIC_DEFAULT_SONNET_MODEL=qwen-local
-# ANTHROPIC_DEFAULT_HAIKU_MODEL=qwen-local
-# ANTHROPIC_DEFAULT_OPUS_MODEL=qwen-local
+# If your OpenAI-compatible backend requires extra auth headers, you can add:
 # ANTHROPIC_CUSTOM_HEADERS=Authorization: Bearer your-token
 
 API_TIMEOUT_MS=3000000

package/dist/src/services/api/claude.js

@@ -201,6 +201,17 @@ import {
   is529Error,
   withRetry
 } from "./withRetry.js";
+function getDefaultTemperature() {
+  const raw = process.env.CLAUDE_CODE_TEMPERATURE;
+  if (raw === undefined) {
+    return 1;
+  }
+  const parsed = Number(raw);
+  if (!Number.isFinite(parsed)) {
+    return 1;
+  }
+  return Math.min(1, Math.max(0, parsed));
+}
 export function getExtraBodyParams(betaHeaders) {
   const extraBodyStr = process.env.CLAUDE_CODE_EXTRA_BODY;
   let result = {};
@@ -861,7 +872,7 @@ ${deferredToolList}
       betasParams.push(cacheEditingBetaHeader);
       logForDebugging("Cache editing beta header enabled for cached microcompact");
     }
-    const temperature = !hasThinking ? options.temperatureOverride ?? 1 : undefined;
+    const temperature = !hasThinking ? options.temperatureOverride ?? getDefaultTemperature() : undefined;
     lastRequestBetas = betasParams;
     return {
       model: normalizeModelStringForAPI(options.model),
@@ -897,7 +908,7 @@ ${deferredToolList}
       logAPIQuery({
         model: options.model,
         messagesLength: logMessagesLength,
-        temperature: options.temperatureOverride ?? 1,
+        temperature: options.temperatureOverride ?? getDefaultTemperature(),
         betas: logBetas,
         permissionMode: permissionContext.mode,
         querySource: options.querySource,

package/dist/src/utils/config.js

@@ -62,7 +62,20 @@ function createDefaultGlobalConfig() {
       approved: [],
       rejected: []
     },
-    env: {},
+    env: {
+      ANTHROPIC_AUTH_TOKEN: "dummy",
+      ANTHROPIC_BASE_URL: "http://127.0.0.1:4000/anthropic",
+      ANTHROPIC_MODEL: "qwen-local",
+      ANTHROPIC_DEFAULT_SONNET_MODEL: "qwen-local",
+      ANTHROPIC_DEFAULT_HAIKU_MODEL: "qwen-local",
+      ANTHROPIC_DEFAULT_OPUS_MODEL: "qwen-local",
+      CLAUDE_CODE_MAX_OUTPUT_TOKENS: "4096",
+      CLAUDE_CODE_MAX_CONTEXT_TOKENS: "65536",
+      CLAUDE_CODE_TEMPERATURE: "0",
+      CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC: "1",
+      CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS: "1",
+      DISABLE_TELEMETRY: "1"
+    },
     tipsHistory: {},
     memoryUsageCount: 0,
     promptQueueUseCount: 0,

package/dist/src/utils/context.js

@@ -26,7 +26,7 @@ export function modelSupports1M(model) {
   return canonical.includes("claude-sonnet-4") || canonical.includes("opus-4-6");
 }
 export function getContextWindowForModel(model, betas) {
-  if (process.env.USER_TYPE === "ant" && process.env.CLAUDE_CODE_MAX_CONTEXT_TOKENS) {
+  if (process.env.CLAUDE_CODE_MAX_CONTEXT_TOKENS) {
     const override = parseInt(process.env.CLAUDE_CODE_MAX_CONTEXT_TOKENS, 10);
     if (!isNaN(override) && override > 0) {
       return override;

package/dist/src/utils/managedEnvConstants.js

@@ -85,11 +85,13 @@ export const SAFE_ENV_VARS = new Set([
   "CLAUDE_CODE_ENABLE_TELEMETRY",
   "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS",
   "CLAUDE_CODE_IDE_SKIP_AUTO_INSTALL",
+  "CLAUDE_CODE_MAX_CONTEXT_TOKENS",
   "CLAUDE_CODE_MAX_OUTPUT_TOKENS",
   "CLAUDE_CODE_SKIP_BEDROCK_AUTH",
   "CLAUDE_CODE_SKIP_FOUNDRY_AUTH",
   "CLAUDE_CODE_SKIP_VERTEX_AUTH",
   "CLAUDE_CODE_SUBAGENT_MODEL",
+  "CLAUDE_CODE_TEMPERATURE",
   "CLAUDE_CODE_USE_BEDROCK",
   "CLAUDE_CODE_USE_FOUNDRY",
   "CLAUDE_CODE_USE_VERTEX",

package/litellm.yaml

@@ -1,9 +1,11 @@
 model_list:
   - model_name: qwen-local
     litellm_params:
-      model: openai/Qwen2.5-Coder-32B-Instruct
-      api_base: http://127.0.0.1:8000/v1
+      model: openai/Qwen3-Next-80B-A3B-Instruct-FP8
+      api_base: http://xinf.lmisp.devops.zgcbank:13002/v1
       api_key: dummy
+      temperature: 0
+      max_tokens: 4096
 
 general_settings:
   master_key: dummy

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "lcclaude",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "private": false,
   "type": "module",
   "imports": {