mocode-ai 0.6.8 → 0.6.9

package/README.md

@@ -32,6 +32,11 @@ MoCode isn't a chat box with a coat of paint — it's an agent that actually get
 - **Optional desktop pet** — A small floating window (`/pet`) shows a stateful character that mirrors agent activity (idle / thinking / tool running / waiting for human). Works as a separate process over WebSocket; quit it with `/pet quit`. Sits beside the terminal, never blocks it.
 - **Slash commands** — `/exit` `/clear` `/context` `/skills` `/compact` `/resume` `/rollback` `/memory` `/reflect` `/init` `/theme` `/model` `/plan` `/auto` `/pet`, with dropdown filtering as you type.
 
+## Documentation
+
+- [中文使用指南](./docs/usage.md) — 菜单式快速上手、命令速查、模式、会话、项目上下文与排障。
+- [Project context](./docs/USAGE_SNAPSHOT_SKILL.md) — Snapshot and Project Skill reference.
+
 ## Installation
 
 Requires Node.js ≥ 18.
@@ -208,14 +213,14 @@ A skill's `description` is injected into the system prompt (progressive disclosu
 
 MoCode uses two complementary systems to help the agent understand your project:
 
-- **Project Snapshot** (automatic + LLM-enhanced): Scans your project files and generates a structured summary including project description, tech stack, key commands, module responsibilities, and directory tree. Built automatically on startup (Phase 1: sync scan ~100ms, Phase 2: async LLM summary ~5s). Stored in `.mocode/snapshot.json`, reused across sessions. Refresh manually with `/snapshot_refresh` after major changes.
-- **Project Skill** (manual + AI-assisted): A knowledge base you maintain capturing insights the agent can't auto-discover — design decisions, architectural patterns, pitfalls, conventions, and development workflow notes. Initialize with `/init` (AI explores your project and drafts the initial content), then edit `.mocode/project-skill.md` to refine. Update via the `project_skill_update` tool during conversations.
+- **Project Snapshot** (automatic + LLM-enhanced, enabled by default): Scans your project files and generates a structured summary including project description, tech stack, key commands, module responsibilities, and directory tree. Stored in `.mocode/snapshot.json` and reused across sessions. Use `/snapshot` to toggle it and `/snapshot_refresh` after major changes.
+- **Project Skill** (manual + AI-assisted, disabled by default): A knowledge base for design decisions, architectural patterns, pitfalls, conventions, and workflow notes. Enable it with `/project_skill on` (or `MOCODE_PROJECT_SKILL=true`), generate an initial draft with `/project_skill init`, then refine `.mocode/project-skill.md`. The agent can update it through `project_skill_update` during conversations.
 
 **Complementary principle**: Snapshot provides *what/where* (facts: files, structure, commands), Skill provides *why/how* (insights: decisions, behaviors, gotchas). No duplication, ~46% token savings compared to the previous approach.
 
-Both are enabled by default. Control via environment variables:
+Control via environment variables:
 - `MOCODE_PROJECT_SNAPSHOT=false` — disable snapshot
-- `MOCODE_PROJECT_SKILL=false` — disable skill
+- `MOCODE_PROJECT_SKILL=true` — enable Project Skill
 
 See [docs/USAGE_SNAPSHOT_SKILL.md](./docs/USAGE_SNAPSHOT_SKILL.md) for detailed usage.
 

package/README.zh-CN.md

@@ -31,6 +31,11 @@ mocode 不是一个套壳聊天框,而是一个能真正动手干活的 agent:
 - **可选桌宠** — 独立悬浮窗(`/pet`)显示一个小角色,镜像 agent 活动(空闲 / 思考 / 跑工具 / 等人工),独立进程走 WebSocket,`/pet quit` 完全关闭。挂在终端外,绝不挡终端。
 - **斜杠命令** — `/exit` `/clear` `/context` `/skills` `/compact` `/resume` `/rollback` `/memory` `/reflect` `/init` `/theme` `/model` `/plan` `/auto` `/pet`,输入时下拉过滤
 
+## 使用文档
+
+- [菜单式使用指南](./docs/usage.md) — 快速上手、命令速查、模式、会话、项目上下文与排障。
+- [项目上下文：Snapshot 与 Project Skill](./docs/USAGE_SNAPSHOT_SKILL.md)
+
 ## 安装
 
 要求 Node.js ≥ 18。

package/dist/agent/core.js

@@ -150,7 +150,7 @@ function pushToolResult(history, tc, output, pruner, lifecycle, _scheduler, runt
  *  这些是纯字符串格式化,无副作用,共享安全。
  */
 export async function runAgentCore(opts) {
-    const { history, userInput, signal, onContextUpdate, hooks, skipRollback } = opts;
+    const { history, userInput, signal, onContextUpdate, hooks } = opts;
     const runtimeContextState = opts.contextState ?? contextState;
     const maxSteps = opts.maxSteps ?? config.maxSteps;
     // 中断还原:LLM 中途可能调 switch_mode 切了模式,abort 时连同模式一起还原回轮首。
@@ -191,7 +191,7 @@ export async function runAgentCore(opts) {
     let savedHistory = history.slice();
     // drop_context 工具的上下文剔除回调:闭包捕获 history,原地剔除无关旧 tool 结果。
     // 保护由 dropContextFromHistory 内部保证:history[0](system)+ 当前轮(最后 user 及其后)永不剔除。
-    // 子 agent 也在自己的 history 上操作(子 agent 独立 history);skipRollback 不影响此行为。
+    // 子 agent 也在自己的 history 上操作(子 agent 独立 history)，文件回滚事务则与主轮共享。
     const dropContext = (filter) => dropContextFromHistory(history, filter);
     // 相关性裁剪 pruner:每个 runAgentCore 实例一个,纯静态、不调 LLM、自动判定 read_file 失效。
     // 开关关闭时为 null,所有 pushToolResult 调用走无 pruner 路径(零行为变化)。
@@ -352,7 +352,7 @@ export async function runAgentCore(opts) {
                         for (const tc of batch)
                             hooks.onToolHeader?.(tc);
                         hooks.onToolStart?.(batch[0].name);
-                        const started = batch.map((tc) => executeTool(tc.name, tc.arguments, signal, { skipRollback, dropContext }));
+                        const started = batch.map((tc) => executeTool(tc.name, tc.arguments, signal, { dropContext }));
                         for (let k = 0; k < batch.length; k++) {
                             const tc = batch[k];
                             const output = await started[k];
@@ -413,7 +413,7 @@ export async function runAgentCore(opts) {
                             i = j;
                             continue; // 全部被拒绝,跳过执行
                         }
-                        const started = allowedBatch.map((tc) => executeTool(tc.name, tc.arguments, signal, { skipRollback, dropContext }));
+                        const started = allowedBatch.map((tc) => executeTool(tc.name, tc.arguments, signal, { dropContext }));
                         // 先批量打印所有头 + 启 spinner(多 task 并发,spinner 只显一个,但 ● 头都打出来)
                         for (const tc of allowedBatch) {
                             hooks.onToolHeader?.(tc);
@@ -468,7 +468,7 @@ export async function runAgentCore(opts) {
                             : null;
                         const { preWriteOld, editStartLine } = readDiffContext(tc, mutationParsed);
                         hooks.onToolStart?.(tc.name);
-                        const output = await executeTool(tc.name, tc.arguments, signal, { skipRollback, dropContext });
+                        const output = await executeTool(tc.name, tc.arguments, signal, { dropContext });
                         hooks.onToolDone?.();
                         hooks.onToolResult?.(tc, output, mutationParsed, preWriteOld, editStartLine);
                         // Thrashing:同上(history 附 hint,UI 干净)

package/dist/agent/index.js

@@ -12,6 +12,8 @@ import { beginTurn } from '../rollback/index.js';
 import { config } from '../config/index.js';
 import { runAgentCore, isMutationTool, } from './core.js';
 import { createPetHooks } from '../pet/state.js';
+import { t } from '../i18n/index.js';
+import { isToolErrorOutput } from '../tools/result.js';
 /** 当前 turn 的 batch id(runAgent 内闭包变量;一条 turn 一轮 tool batch 结束即清空)。 */
 let currentBatchId = null;
 /** 取 userInput 的首行:字符串直接 split;多模态 parts 找首个 text part 再 split。 */
@@ -40,7 +42,7 @@ function writeToolResult(tc, output, parsed, preWriteOld, editStartLine) {
     if (!currentBatchId)
         return;
     let diff = null;
-    if (isMutationTool(tc.name) && parsed && !output.startsWith('错误')) {
+    if (isMutationTool(tc.name) && parsed && !isToolErrorOutput(output)) {
         diff = renderFileChange({
             path: String(parsed.path ?? ''),
             kind: tc.name === 'edit_file' ? 'edit' : 'write',
@@ -143,9 +145,9 @@ onContextUpdate) {
                 hasPendingTextBoundary = false;
             }
             if (name)
-                spinner.start(`生成 ${name}`);
+                spinner.start(t('agent.generating', { tool: name }));
         },
-        onStepStart: () => spinner.start('思考中'),
+        onStepStart: () => spinner.start(t('agent.thinking')),
         onChatDone: () => spinner.stop(),
         onTextEnd: () => {
             if (lastChar && lastChar !== '\n') {
@@ -164,7 +166,7 @@ onContextUpdate) {
             toolBatchFollowsText = false;
             writeToolHeader(tc);
         },
-        onToolStart: (name) => spinner.start(`执行 ${name}`),
+        onToolStart: (name) => spinner.start(t('agent.executing', { tool: name })),
         onToolDone: () => spinner.stop(),
         onToolResult: (tc, output, parsed, preWriteOld, editStartLine) => writeToolResult(tc, output, parsed, preWriteOld, editStartLine),
         onToolBatchEnd: () => {
@@ -172,23 +174,23 @@ onContextUpdate) {
         },
         onNoReply: () => {
             flushToolBatch();
-            layout.contentWrite(`${ui.dim}(无回复)${ui.reset}\n`);
+            layout.contentWrite(`${ui.dim}${t('agent.noReply')}${ui.reset}\n`);
         },
         onMaxSteps: () => {
             flushToolBatch();
-            layout.contentWrite(`  ${ui.yellow}●${ui.reset} ${ui.yellow}达到最大步数(${config.maxSteps}),本轮停止。${ui.reset}\n`);
+            layout.contentWrite(`  ${ui.yellow}●${ui.reset} ${ui.yellow}${t('agent.maxSteps', { count: config.maxSteps })}${ui.reset}\n`);
         },
         onAbort: () => {
             spinner.stop();
             if (lastChar && lastChar !== '\n')
                 layout.contentWrite('\n');
             flushToolBatch();
-            layout.contentWrite(`${ui.dim}(已中断)${ui.reset}\n`);
+            layout.contentWrite(`${ui.dim}${t('agent.aborted')}${ui.reset}\n`);
         },
         onDone: (elapsedMs, usage) => {
             flushToolBatch();
             const tok = formatTurnTokens(usage);
-            layout.contentWrite(`  ${ui.dim}✻ Worked for ${fmtElapsed(elapsedMs)}${tok}${ui.reset}\n`);
+            layout.contentWrite(`  ${ui.dim}✻ ${t('agent.workedFor', { elapsed: fmtElapsed(elapsedMs) })}${tok}${ui.reset}\n`);
             // 内容区触底时，DECSTBM 增量滚屏可能只推进物理终端，未把 Worked 前已在
             // buffer 中的空行完整画出来；用户滚动/点击触发 repaint 后才“突然”出现。
             // 轮次收尾立即按 buffer 原子重画，使未满屏与触底滚屏的布局一致。

package/dist/agent/spawn.js

@@ -7,13 +7,10 @@
 //  - 独立 history:不共享主对话,避免子任务的工具噪声污染主上下文。
 //  - 系统提示复用主 agent 组装链(config.systemPrompt + memory 段 + skills 段)+ 子 agent 角色后缀。
 //  - 工具子集:按白名单从 chatTools 过滤;无白名单 = 全量(但 task 工具调用方通常会限定只读)。
-//  - 不调 beginTurn(不进主回滚链);skipRollback=true 使文件 mutation 跳过 recordMutation,
-//    子 agent 的改动不进主回滚快照链(靠 git 兜底,见下方 skipRollback 注释)。
+//  - 不调 beginTurn：子 agent 共享主 agent 当前轮次；其文件修改进入同一回滚事务。
 //  - 步数上限默认更低(config.subAgentMaxSteps ?? 50),防子任务失控耗尽配额。
 //  - 中断透传:opts.signal(主 agent 的 abort signal)透传给 runAgentCore → chat/executeTool,
 //    主 Ctrl+C 树杀子 agent(chat 流式 abort + run_command/web_fetch 即时取消)。
-//  - 逻辑隔离(回滚):skipRollback=true,子 agent 的 write_file/edit_file 改动不进主回滚快照链,
-//    主 /rollback 不撤销子 agent 改动(靠 git 兜底)。子 agent 与主 agent 共享 cwd(文件改动可见)。
 import { chatTools } from '../llm/index.js';
 import { config, isMemoryEnabled } from '../config/index.js';
 import { effectiveSystemPrompt } from '../skills/index.js';
@@ -126,7 +123,6 @@ export async function spawnAgent(opts) {
         hooks,
         maxSteps,
         toolsOverride,
-        skipRollback: true, // 逻辑隔离:子 agent 文件改动不进主回滚快照链,主 /rollback 不撤销(靠 git 兜底)
         contextState: localContextState,
     });
     return {

package/dist/commands/config.js

@@ -1,5 +1,6 @@
 import * as readline from 'node:readline';
 import { CONFIG_PATH, readConfigFile, writeConfigKeys } from '../config/file.js';
+import { detectLanguage, setLanguage, t } from '../i18n/index.js';
 function ask(rl, q) {
     return new Promise((resolve) => {
         rl.question(q, (ans) => resolve(ans.trim()));
@@ -14,23 +15,24 @@ function ask(rl, q) {
  */
 export async function runConfigWizard() {
     const cur = readConfigFile();
+    setLanguage(detectLanguage(process.env.MOCODE_LANGUAGE ?? cur.MOCODE_LANGUAGE));
     const rl = readline.createInterface({ input: process.stdin, output: process.stdout });
-    console.log(`mocode 配置向导 — 写入 ${CONFIG_PATH}(Ctrl+C 取消)\n`);
-    const baseURLIn = await ask(rl, `LLM_BASE_URL${cur.LLM_BASE_URL ? ` [${cur.LLM_BASE_URL}]` : ''}(如 https://open.bigmodel.cn/api/v3): `);
+    console.log(`${t('config.title', { path: CONFIG_PATH })}\n`);
+    const baseURLIn = await ask(rl, `LLM_BASE_URL${cur.LLM_BASE_URL ? ` [${cur.LLM_BASE_URL}]` : ''}${t('config.example')}: `);
     const baseURL = baseURLIn || cur.LLM_BASE_URL || '';
     if (!baseURL) {
-        console.error('\n[config] LLM_BASE_URL 不能为空,已取消。');
+        console.error(t('config.required', { key: 'LLM_BASE_URL' }));
         rl.close();
         process.exit(1);
     }
-    const apiKeyIn = await ask(rl, `LLM_API_KEY${cur.LLM_API_KEY ? ' [已设置,回车保留]' : ''}: `);
+    const apiKeyIn = await ask(rl, `LLM_API_KEY${cur.LLM_API_KEY ? t('config.keySet') : ''}: `);
     const apiKey = apiKeyIn || cur.LLM_API_KEY || '';
     if (!apiKey) {
-        console.error('\n[config] LLM_API_KEY 不能为空,已取消。');
+        console.error(t('config.required', { key: 'LLM_API_KEY' }));
         rl.close();
         process.exit(1);
     }
-    const modelIn = await ask(rl, `LLM_MODEL${cur.LLM_MODEL ? ` [${cur.LLM_MODEL}]` : ''}(回车默认 gpt-4o-mini): `);
+    const modelIn = await ask(rl, `LLM_MODEL${cur.LLM_MODEL ? ` [${cur.LLM_MODEL}]` : ''}${t('config.modelDefault')}: `);
     const model = modelIn || cur.LLM_MODEL || 'gpt-4o-mini';
     rl.close();
     // 合并:保留其它键,只覆盖三键。
@@ -39,5 +41,5 @@ export async function runConfigWizard() {
         LLM_API_KEY: apiKey,
         LLM_MODEL: model,
     });
-    console.log(`\n已写入 ${CONFIG_PATH}。现在运行 \`mocode\` 即可启动(任意目录、任意终端)。`);
+    console.log(t('config.done', { path: CONFIG_PATH }));
 }

package/dist/config/index.js

@@ -6,6 +6,7 @@ import { loadSnapshot } from '../project-snapshot/index.js';
 import { buildProjectSkillSection } from '../project-skill/index.js';
 import { getSandboxRoot } from '../sandbox/root.js';
 import { getCurrentSessionId } from '../session/state.js';
+import { detectLanguage, setLanguage, t, } from '../i18n/index.js';
 /**
  * 按优先级加载配置文件并回填 process.env:
  *   候选(后者覆盖前者,优先级升序):<cwd>/.env(兼容旧用法,最低)→ ~/.mocode/config(全局)→ <cwd>/.mocode/config(项目级覆盖,最高)。
@@ -34,14 +35,16 @@ function loadEnvFiles() {
             process.env[k] = v;
     }
 }
-// 在 loadEnvFiles 回填前捕获:MOCODE_THEME 是否由 shell 设置(决定 /theme 写文件是否下次启动生效)。
+// 在 loadEnvFiles 回填前捕获:MOCODE_THEME / MOCODE_LANGUAGE 是否由 shell 设置。
 const themeFromShell = process.env.MOCODE_THEME !== undefined;
+export const languageFromShell = process.env.MOCODE_LANGUAGE !== undefined;
 // 在 loadEnvFiles 回填前捕获:哪些 LLM 键由 shell 设置(决定 /model 写文件是否下次启动生效)。
 // 仿 themeFromShell 模式:shell export 的环境变量在 loadEnvFiles 中不被回填(优先级最高),
 // 故 /model 写入 ~/.mocode/config 的同名键下次启动会被 shell 值覆盖——据此给 dim 警告。
 const LLM_ENV_KEYS = ['LLM_BASE_URL', 'LLM_API_KEY', 'LLM_MODEL', 'CONTEXT_WINDOW_TOKENS'];
 const llmKeysFromShell = LLM_ENV_KEYS.filter((k) => process.env[k] !== undefined);
 loadEnvFiles();
+setLanguage(detectLanguage(process.env.MOCODE_LANGUAGE));
 /**
  * 取环境变量;缺则返回空字符串(不退出)。
  * 历史上缺 LLM_BASE_URL/LLM_API_KEY 会 process.exit(1),但 /model 命令已能在 REPL 内配置模型,
@@ -163,14 +166,14 @@ const SYSTEM_PROMPT_MEMORY_SECTION = `
  * 句都不出现,且 read-only 列表里的 memory_search/memory_list 也移除——避免提示词里出现
  * 根本不存在的工具名引起 LLM 调不到。
  */
-const PLAN_RESEARCH_RULES = `
-- Research enough to locate relevant code, trace call paths, and understand existing conventions. Use the codegraph-first Workflow, but do not repeat information already retrieved in this session.
-- Produce an actionable plan: files and reasons, ordered steps, edge cases, and verification (typecheck / tests / build).
-- When ready, MUST call the \`ask_human\` tool with a concise summary and exactly these options:
-  1. "按计划执行 (switch to auto and implement)" — call \`switch_mode("auto")\` in the same turn, then implement.
-  2. "继续细化方案 (stay in plan, refine)" — remain in plan and refine.
-  3. "取消 / 暂不执行 (abort)" — stop without switching mode.
-- Never silently switch or stop. Do not ask approval in plain text; \`ask_human\` is the approval channel.
+const PLAN_RESEARCH_RULES = `
+- Research enough to locate relevant code, trace call paths, and understand existing conventions. Use the codegraph-first Workflow, but do not repeat information already retrieved in this session.
+- Produce an actionable plan: files and reasons, ordered steps, edge cases, and verification (typecheck / tests / build).
+- When ready, MUST call the \`ask_human\` tool with a concise summary and exactly these options:
+  1. "按计划执行 (switch to auto and implement)" — call \`switch_mode("auto")\` in the same turn, then implement.
+  2. "继续细化方案 (stay in plan, refine)" — remain in plan and refine.
+  3. "取消 / 暂不执行 (abort)" — stop without switching mode.
+- Never silently switch or stop. Do not ask approval in plain text; \`ask_human\` is the approval channel.
 - The REPL approval prompt is only a fallback; do not rely on it.`;
 function buildPlanModeSuffix() {
     const memoryTools = isMemoryEnabled()
@@ -182,11 +185,11 @@ function buildPlanModeSuffix() {
     const removed = ['write_file', 'edit_file', 'run_command', memoryTools]
         .filter(Boolean)
         .join(', ');
-    return `
-
-## ⛯ PLAN MODE (active now)
-You are in PLAN mode: investigate and design only — do NOT execute or change anything.
-- Removed from your tool list: ${removed}. Use only these read-only tools: ${readOnlyTools}.
+    return `
+
+## ⛯ PLAN MODE (active now)
+You are in PLAN mode: investigate and design only — do NOT execute or change anything.
+- Removed from your tool list: ${removed}. Use only these read-only tools: ${readOnlyTools}.
 ${PLAN_RESEARCH_RULES}`;
 }
 /** 兼容旧名字:repl 的 buildSystemMessage 仍引 PLAN_MODE_SUFFIX(变量)。运行时按需现拼。 */
@@ -198,8 +201,8 @@ export function buildBasePrompt() {
     const planLine = isMemoryEnabled()
         ? '- For complex or multi-step tasks, the user may switch to PLAN mode (Shift+Tab): your editing/command/memory-write tools are then removed from your tool list, and you must research with read-only tools only and produce a step-by-step plan (no execution). On approval the session returns to auto mode to execute the plan.'
         : '- For complex or multi-step tasks, the user may switch to PLAN mode (Shift+Tab): your editing/command tools are then removed from your tool list, and you must research with read-only tools only and produce a step-by-step plan (no execution). On approval the session returns to auto mode to execute the plan.';
-    return `## Core behavior
-You are mocode, a terminal coding agent. Complete programming tasks through a "think → call tool → observe result → think again" loop until solved. Reply to the user in Chinese.
+    return `## Core behavior
+You are mocode, a terminal coding agent. Complete programming tasks through a "think → call tool → observe result → think again" loop until solved. ${t('assistant.languageInstruction')}
 
 ## 模式 (Modes)
 ${autoAllToolsLine}
@@ -207,31 +210,31 @@ ${planLine}
 
 ${PLATFORM_NOTE}
 
-## Tool details
-### Token-efficient execution
-- First check whether the answer is already in this conversation or a previous tool result. If yes, answer directly; do not re-run tools "to be safe".
-- Plan the complete sub-task before calling tools. Batch independent reads in one turn. Because tool calls in one response execute without intermediate model reasoning, never batch a read with an edit that depends on its result.
-- Do not read "just to see". Read only what supports the next decision. Re-read after a change, compaction, stale state, or uncertain line context.
-- Prefer one precise call over overlapping searches. If a call fails, inspect the error and change the approach instead of repeating it unchanged.
-- Batch only independent read-only calls. After their results arrive, make the dependent edit in the next turn; then batch independent edits and one final verification when their exact inputs are already known.
-- Read only what supports the next decision; verify once after a related edit set, not after every edit.
-- Do not repeat an unchanged failing call; after three unproductive attempts, change tools or ask for the missing decision.
-- For \`edit_file\`, derive \`old_string\` by copying the exact relevant lines from the latest successful \`read_file\` of that same path; never reconstruct it from memory, a summary, grep output, or a previous diff. That read becomes stale after any edit/write to the path, compaction/resume, or a possible external change. On an \`old_string\` mismatch, re-read the exact region and retry once with the newly returned text; never retry identical arguments.
+## Tool details
+### Token-efficient execution
+- First check whether the answer is already in this conversation or a previous tool result. If yes, answer directly; do not re-run tools "to be safe".
+- Plan the complete sub-task before calling tools. Batch independent reads in one turn. Because tool calls in one response execute without intermediate model reasoning, never batch a read with an edit that depends on its result.
+- Do not read "just to see". Read only what supports the next decision. Re-read after a change, compaction, stale state, or uncertain line context.
+- Prefer one precise call over overlapping searches. If a call fails, inspect the error and change the approach instead of repeating it unchanged.
+- Batch only independent read-only calls. After their results arrive, make the dependent edit in the next turn; then batch independent edits and one final verification when their exact inputs are already known.
+- Read only what supports the next decision; verify once after a related edit set, not after every edit.
+- Do not repeat an unchanged failing call; after three unproductive attempts, change tools or ask for the missing decision.
+- For \`edit_file\`, derive \`old_string\` by copying the exact relevant lines from the latest successful \`read_file\` of that same path; never reconstruct it from memory, a summary, grep output, or a previous diff. That read becomes stale after any edit/write to the path, compaction/resume, or a possible external change. On an \`old_string\` mismatch, re-read the exact region and retry once with the newly returned text; never retry identical arguments.
 
-## Workflow
-- Understand requirements and current code before acting; do not guess.
-- If \`.codegraph/\` exists, use \`codegraph\` first for unfamiliar code questions. Use direct reads for known or recently changed files.
-- After modifications, run the smallest relevant verification, then typecheck/build when appropriate. Never claim success without evidence.
-- Use web search only when freshness materially affects the answer (new APIs, versions, security, current UI conventions).
+## Workflow
+- Understand requirements and current code before acting; do not guess.
+- If \`.codegraph/\` exists, use \`codegraph\` first for unfamiliar code questions. Use direct reads for known or recently changed files.
+- After modifications, run the smallest relevant verification, then typecheck/build when appropriate. Never claim success without evidence.
+- Use web search only when freshness materially affects the answer (new APIs, versions, security, current UI conventions).
 
-## Tool rules
-- Precise path/symbol → go directly to \`read_file\` or \`codegraph node\`; use \`glob\`/\`grep\` only for discovery.
-- Before editing, read the exact target region and use its verbatim text as \`old_string\`. Use \`edit_file\` for unique local replacements and \`write_file\` for new/full files.
-- Local edits require an exact unique match; use \`write_file\` for new/full files.
-- Use \`glob\`/\`grep\` for discovery and \`run_command\` for execution or verification, not file existence checks. State intent before side effects.
-- Call \`ask_human\` only when a real user decision is required; otherwise decide and proceed.
-- Drop stale tool output when it no longer supports the current sub-task. Keep only evidence needed for the next decision.
-- Batch independent writes only when each input is already known and their order does not matter. Keep dependent mutations sequential. Combine a clear shell workflow in one command; follow up when its result creates a decision.
+## Tool rules
+- Precise path/symbol → go directly to \`read_file\` or \`codegraph node\`; use \`glob\`/\`grep\` only for discovery.
+- Before editing, read the exact target region and use its verbatim text as \`old_string\`. Use \`edit_file\` for unique local replacements and \`write_file\` for new/full files.
+- Local edits require an exact unique match; use \`write_file\` for new/full files.
+- Use \`glob\`/\`grep\` for discovery and \`run_command\` for execution or verification, not file existence checks. State intent before side effects.
+- Call \`ask_human\` only when a real user decision is required; otherwise decide and proceed.
+- Drop stale tool output when it no longer supports the current sub-task. Keep only evidence needed for the next decision.
+- Batch independent writes only when each input is already known and their order does not matter. Keep dependent mutations sequential. Combine a clear shell workflow in one command; follow up when its result creates a decision.
 
 ## Large file writes (avoid token-cap truncation)
 - \`write_file\` / \`edit_file\` arguments are part of the model's JSON output — a single tool call's content > ~5K tokens risks mid-stream truncation when the model's max output (default 8K–16K tokens) is exceeded, producing a "arguments 不是合法 JSON" error. Even with \`MAX_TOKENS=32000\` set, huge files still risk truncation.
@@ -248,8 +251,8 @@ ${PLATFORM_NOTE}
 - Confirm with the user before irreversible or outward-facing operations (delete, overwrite existing files, push, request external services), unless explicitly authorized.
 - Operate only within authorized scope; when unsure, ask — don't guess.
 
-## Project context (dynamic reference)
-${buildSnapshotSection()}${config.projectSkillEnabled ? buildProjectSkillSection() : ''}${memorySection}${buildNotepadSection()}
+## Project context (dynamic reference)
+${buildSnapshotSection()}${config.projectSkillEnabled ? buildProjectSkillSection() : ''}${memorySection}${buildNotepadSection()}
 
 ## Session Notepad — working notes file
 ${getCurrentSessionId()
@@ -445,3 +448,8 @@ export function updateSnapshotConfig(enabled) {
     config.projectSnapshotEnabled = enabled;
     process.env.MOCODE_PROJECT_SNAPSHOT = enabled ? 'true' : 'false';
 }
+/** 切换界面与模型回复语言；持久化由 REPL 调用 config/file.ts 完成。 */
+export function updateLanguageConfig(language) {
+    setLanguage(language);
+    process.env.MOCODE_LANGUAGE = language;
+}