foliko 2.0.20 → 2.0.22

package/src/agent/chat.js

@@ -19,7 +19,7 @@ const {
 } = require('ai');
 const { z } = require('zod');
 const { autoSplitToolResult } = require('../../plugins/executors/data-splitter');
-const { normalizeToolOutputs, validateAll } = require('../utils/message-validator');
+const { normalizeToolOutputs, validateAll, sanitizeForLLM } = require('../utils/message-validator');
 const { cleanResponse } = require('../utils');
 const { ChatQueueManager } = require('../common/queue');
 const { TokenCounter } = require('../llm/tokens');
@@ -1073,6 +1073,25 @@ class AgentChatHandler extends EventEmitter {
     this._repairInvalidToolCallsInMessages(messages);
     this._repairAndPrefixHistoricalIds(messages);
 
+    // ★ 关键：去重重复的 tool-call ID
+    // 场景：DeepSeek API 在某些情况下会为不同的 LLM 调用生成相同的 toolCallId
+    //   （或者重放历史消息时保留了原始 ID）。同一请求中 tool-call ID 必须唯一，
+    //   否则 DeepSeek 会抛 "invalid params, tool result's tool id(X) not found" 错误。
+    // 策略：保留第一次出现的 ID，后续出现重命名为 <id>_v2/_v3... 并同步更新 tool-result。
+    const deduped = this._dedupeToolCallIds(messages);
+    if (deduped > 0) {
+      logger.warn(`[_prepareSession] 去重 ${deduped} 个重复 tool-call ID`);
+    }
+
+    // ★ 防御性 sanitize：确保所有 message 符合 AI SDK ModelMessage[] schema
+    // 修复/丢弃：content 类型错误、part 缺 type、tool-result.output 异常等
+    const sanitized = sanitizeForLLM(messages);
+    if (sanitized.fixed > 0 || sanitized.dropped > 0) {
+      logger.warn(`[_prepareSession] sanitize: fixed=${sanitized.fixed}, dropped=${sanitized.dropped}`);
+      messages.length = 0;
+      messages.push(...sanitized.messages);
+    }
+
     // 单次遍历完成所有验证 + 规范化（validateAll 内部处理 error-text/error-json）
     const validated = validateAll(messages);
     if (validated.length !== messages.length) {
@@ -1106,6 +1125,85 @@ class AgentChatHandler extends EventEmitter {
     return { messageStore, messages, userMessage };
   }
 
+  /**
+   * ★ 去重 messages 中重复的 tool-call / tool-result ID
+   * DeepSeek API 严格要求同一请求中 toolCallId 唯一。
+   * 当历史会话中多个 assistant 消息复用了同一 toolCallId（DeepSeek 偶尔会这样生成），
+   * 第二个 tool-result 会找不到匹配 → 报 "tool result's tool id not found"。
+   *
+   * 策略：保留第一次出现的 ID；后续出现重命名为 `<id>_v2`、`<id>_v3`...，
+   *       并同步更新紧随其后的 tool-result ID（保留 call/result 配对）。
+   *
+   * @param {Array} messages 消息数组（原地修改）
+   * @returns {number} 修复的 ID 数量
+   */
+  _dedupeToolCallIds(messages) {
+    if (!Array.isArray(messages)) return 0;
+
+    let fixedCount = 0;
+    // 跟踪第一次出现某 ID 的位置（用于同步 result 改名）
+    const firstSeen = new Map(); // id → {msgIdx, itemIdx, kind: 'call'|'result'}
+
+    for (let msgIdx = 0; msgIdx < messages.length; msgIdx++) {
+      const msg = messages[msgIdx];
+      if (msg.role === 'assistant' && Array.isArray(msg.content)) {
+        for (let itemIdx = 0; itemIdx < msg.content.length; itemIdx++) {
+          const item = msg.content[itemIdx];
+          if (!item) continue;
+          if ((item.type === 'tool-call' || item.type === 'tool-use') && item.toolCallId) {
+            const id = item.toolCallId;
+            if (firstSeen.has(id)) {
+              // 冲突：生成唯一新 ID（不依赖全局计数器，避免跨调用泄露）
+              const newId = `${id}_dup${msgIdx}_${itemIdx}`;
+              item.toolCallId = newId;
+              // 同步修改紧随其后的 tool-result
+              const renamed = this._renameFollowingToolResult(messages, msgIdx, id, newId);
+              fixedCount += 1 + (renamed ? 1 : 0);
+              // 不更新 firstSeen（原始 ID 已被第一次声明占用）
+            } else {
+              firstSeen.set(id, { msgIdx, itemIdx, kind: 'call' });
+            }
+          }
+        }
+      } else if (msg.role === 'tool' && Array.isArray(msg.content)) {
+        for (let itemIdx = 0; itemIdx < msg.content.length; itemIdx++) {
+          const item = msg.content[itemIdx];
+          if (!item) continue;
+          if ((item.type === 'tool-result' || item.type === 'tool_result') && item.toolCallId) {
+            const id = item.toolCallId;
+            // 记录 tool-result 第一次出现（不与 call 比较，因为可能已被上面的分支处理过）
+            if (!firstSeen.has(id)) {
+              firstSeen.set(id, { msgIdx, itemIdx, kind: 'result' });
+            }
+          }
+        }
+      }
+    }
+    return fixedCount;
+  }
+
+  /**
+   * 在 msgIdx 之后的 tool 消息中，找到第一个 toolCallId === oldId 的项，改名为 newId
+   * 遇到下一个 assistant 消息时停止（确保只改当前 assistant 的 result）
+   * @private
+   */
+  _renameFollowingToolResult(messages, afterMsgIdx, oldId, newId) {
+    for (let i = afterMsgIdx + 1; i < messages.length; i++) {
+      const m = messages[i];
+      if (m.role === 'assistant') return false;
+      if (m.role === 'tool' && Array.isArray(m.content)) {
+        for (const item of m.content) {
+          if (!item) continue;
+          if ((item.type === 'tool-result' || item.type === 'tool_result') && item.toolCallId === oldId) {
+            item.toolCallId = newId;
+            return true;
+          }
+        }
+      }
+    }
+    return false;
+  }
+
   /**
    * 判断两个 user content 是否等价（用于末尾 user 去重）
    * @private
@@ -1375,6 +1473,15 @@ class AgentChatHandler extends EventEmitter {
         const repairedCount = this._repairInvalidToolCallsInMessages(inputMessages);
         this._repairAndPrefixHistoricalIds(inputMessages);
 
+        // ★ 防御性 sanitize：确保所有 message 符合 AI SDK ModelMessage[] schema
+        // 每次 LLM step 前都跑一次（messages 数组里可能有 AI SDK 上一步加进来的）
+        const sanitized = sanitizeForLLM(inputMessages);
+        if (sanitized.fixed > 0 || sanitized.dropped > 0) {
+          logger.warn(`[PrepareStep] sanitize: fixed=${sanitized.fixed}, dropped=${sanitized.dropped}`);
+          inputMessages.length = 0;
+          inputMessages.push(...sanitized.messages);
+        }
+
         // 单次遍历完成所有验证 + 规范化
         const validated = validateAll(inputMessages);
         if (validated.length !== inputMessages.length) {

package/src/agent/sub.js

@@ -30,7 +30,7 @@ class SubAgent extends BaseAgent {
     this.defaulTools = ['ext_skill', 'skill_load', 'ext_call', 'subagent_call'];
     this.bindTools = {
       read: 'read_file', write: 'write_file', edit: 'modify_file',
-      glob: 'read_directory', grep: 'search_file', bash: 'shell_exec',
+      glob: 'read_directory', grep: 'search', bash: 'shell_exec',
     };
 
     this._customSystemPrompt = config.systemPrompt || null;

package/src/tool/router.js

@@ -15,7 +15,7 @@ const INTENT_PATTERNS = {
     ],
     tools: [
       'read_file', 'write_file', 'delete_file', 'modify_file',
-      'read_directory', 'create_directory', 'search_file',
+      'read_directory', 'create_directory', 'search',
     ],
   },
   code_development: {
@@ -34,7 +34,7 @@ const INTENT_PATTERNS = {
   },
   data_analysis: {
     keywords: ['分析', '数据', '统计', '查询', 'analyze', 'data', 'statistics', 'query'],
-    tools: ['execute_command', 'py_execute', 'search_file'],
+    tools: ['execute_command', 'py_execute', 'search', 'read_file', 'write_file'],
   },
   plugin_management: {
     keywords: ['插件', '重载', '加载插件', 'plugin', 'reload', 'load plugin'],

package/src/utils/message-validator.js

@@ -499,6 +499,191 @@ function filterPairedMessages(messages, keepRecentAssistant = 3) {
     .map((i) => messages[i]);
 }
 
+/**
+ * ★ 防御性 sanitize：确保 messages 数组符合 AI SDK 的 ModelMessage[] schema
+ *
+ * 已知会失败的情况：
+ * 1. `user.content` 不是 string 或 array（如 null、undefined、object）
+ * 2. `assistant.content` 不是 string 或 array
+ * 3. `tool.content` 不是 array
+ * 4. user/assistant content 是 array，但 part 缺少 `type` 字段或 type 不在白名单
+ * 5. tool content 是 array，但 part 不是 tool-result/tool-approval-response
+ * 6. tool-result 的 `output` 缺少 type 或 type 不在 AI SDK 允许集
+ *
+ * 策略：
+ * - 可修复 → 修复（log warn）
+ * - 不可修复 → 丢弃整条消息（log warn）
+ *
+ * @param {Array} messages 消息数组
+ * @returns {{ messages: Array, fixed: number, dropped: number }}
+ */
+function sanitizeForLLM(messages) {
+  if (!Array.isArray(messages)) return { messages: [], fixed: 0, dropped: 0 };
+  const VALID_USER_PARTS = new Set(['text', 'image', 'file']);
+  const VALID_ASSISTANT_PARTS = new Set(['text', 'file', 'reasoning', 'tool-call', 'tool-result', 'tool-approval-request']);
+  const VALID_TOOL_PARTS = new Set(['tool-result', 'tool-approval-response']);
+  const VALID_OUTPUT_TYPES = new Set(['text', 'json', 'execution-denied', 'content', 'error-text', 'error-json']);
+
+  let fixed = 0;
+  let dropped = 0;
+  const out = [];
+
+  for (let i = 0; i < messages.length; i++) {
+    const m = messages[i];
+    if (!m || typeof m !== 'object' || !m.role) {
+      console.warn(`[sanitizeForLLM] dropped msg[${i}]: missing role`);
+      dropped++;
+      continue;
+    }
+
+    if (m.role === 'system') {
+      // system: content 必须是 string
+      if (typeof m.content !== 'string') {
+        console.warn(`[sanitizeForLLM] fixed system msg[${i}]: content ${typeof m.content} → string`);
+        out.push({ ...m, content: String(m.content ?? '') });
+        fixed++;
+      } else {
+        out.push(m);
+      }
+      continue;
+    }
+
+    if (m.role === 'user') {
+      // user: content 必须是 string 或 array（text/image/file parts）
+      if (typeof m.content === 'string') {
+        out.push(m);
+        continue;
+      }
+      if (Array.isArray(m.content)) {
+        const newContent = [];
+        for (const part of m.content) {
+          if (part && typeof part === 'object' && VALID_USER_PARTS.has(part.type)) {
+            newContent.push(part);
+          } else {
+            console.warn(`[sanitizeForLLM] dropped user msg[${i}] part: ${JSON.stringify(part).slice(0, 100)}`);
+            fixed++;
+          }
+        }
+        if (newContent.length === 0) {
+          console.warn(`[sanitizeForLLM] dropped user msg[${i}]: all parts invalid`);
+          dropped++;
+          continue;
+        }
+        out.push({ ...m, content: newContent });
+        continue;
+      }
+      // content 是 object/null/undefined → 转为 string 或丢弃
+      if (m.content == null) {
+        console.warn(`[sanitizeForLLM] dropped user msg[${i}]: null content`);
+        dropped++;
+        continue;
+      }
+      // 转为字符串
+      try {
+        out.push({ ...m, content: JSON.stringify(m.content) });
+        fixed++;
+        console.warn(`[sanitizeForLLM] fixed user msg[${i}]: content object → string`);
+      } catch {
+        out.push({ ...m, content: String(m.content) });
+        fixed++;
+      }
+      continue;
+    }
+
+    if (m.role === 'assistant') {
+      // assistant: content 必须是 string 或 array
+      if (typeof m.content === 'string') {
+        out.push(m);
+        continue;
+      }
+      if (Array.isArray(m.content)) {
+        const newContent = [];
+        for (const part of m.content) {
+          if (!part || typeof part !== 'object' || !part.type) {
+            console.warn(`[sanitizeForLLM] dropped assistant msg[${i}] part: no type - ${JSON.stringify(part).slice(0, 100)}`);
+            fixed++;
+            continue;
+          }
+          if (VALID_ASSISTANT_PARTS.has(part.type)) {
+            newContent.push(part);
+          } else {
+            console.warn(`[sanitizeForLLM] dropped assistant msg[${i}] part: unknown type "${part.type}"`);
+            fixed++;
+          }
+        }
+        if (newContent.length === 0) {
+          console.warn(`[sanitizeForLLM] dropped assistant msg[${i}]: all parts invalid`);
+          dropped++;
+          continue;
+        }
+        out.push({ ...m, content: newContent });
+        continue;
+      }
+      if (m.content == null) {
+        console.warn(`[sanitizeForLLM] dropped assistant msg[${i}]: null content`);
+        dropped++;
+        continue;
+      }
+      try {
+        out.push({ ...m, content: JSON.stringify(m.content) });
+        fixed++;
+        console.warn(`[sanitizeForLLM] fixed assistant msg[${i}]: content ${typeof m.content} → string`);
+      } catch {
+        out.push({ ...m, content: String(m.content) });
+        fixed++;
+      }
+      continue;
+    }
+
+    if (m.role === 'tool') {
+      // tool: content 必须是 array (tool-result 或 tool-approval-response)
+      if (!Array.isArray(m.content)) {
+        console.warn(`[sanitizeForLLM] dropped tool msg[${i}]: content is ${typeof m.content}, not array`);
+        dropped++;
+        continue;
+      }
+      const newContent = [];
+      for (const part of m.content) {
+        if (!part || typeof part !== 'object' || !part.type) {
+          console.warn(`[sanitizeForLLM] dropped tool msg[${i}] part: no type`);
+          fixed++;
+          continue;
+        }
+        if (!VALID_TOOL_PARTS.has(part.type)) {
+          console.warn(`[sanitizeForLLM] dropped tool msg[${i}] part: type "${part.type}" not in tool union`);
+          fixed++;
+          continue;
+        }
+        // 修复 tool-result.output
+        if (part.type === 'tool-result') {
+          if (!part.output || typeof part.output !== 'object' || !VALID_OUTPUT_TYPES.has(part.output.type)) {
+            console.warn(`[sanitizeForLLM] fixed tool msg[${i}] tool-result: bad output - ${JSON.stringify(part.output).slice(0, 100)}`);
+            const out2 = (part.output == null) ? { type: 'text', value: String(part.output ?? '') } :
+              { type: 'text', value: JSON.stringify(part.output) };
+            newContent.push({ ...part, output: out2 });
+            fixed++;
+            continue;
+          }
+        }
+        newContent.push(part);
+      }
+      if (newContent.length === 0) {
+        console.warn(`[sanitizeForLLM] dropped tool msg[${i}]: all parts invalid`);
+        dropped++;
+        continue;
+      }
+      out.push({ ...m, content: newContent });
+      continue;
+    }
+
+    // 未知 role：丢弃
+    console.warn(`[sanitizeForLLM] dropped msg[${i}]: unknown role "${m.role}"`);
+    dropped++;
+  }
+
+  return { messages: out, fixed, dropped };
+}
+
 module.exports = {
   validateMessagesPairing,
   validateToolCalls,
@@ -506,4 +691,5 @@ module.exports = {
   normalizeToolOutputs,
   filterOrphanToolCalls,
   filterPairedMessages,
+  sanitizeForLLM,
 };

package/tests/core/chat-tool.test.js

@@ -0,0 +1,187 @@
+import { describe, it, expect } from 'vitest';
+
+// 测试 file-system 插件的 chat 工具修复
+// 通过读源码 + 模拟 execute 来验证行为
+
+const fs = require('fs');
+const path = require('path');
+
+const FileSystemPlugin = require('../../plugins/io/file-system/index.js');
+
+function makeFakeFw(opts = {}) {
+  const llmConfig = opts.llmConfig || { model: 'test-model', provider: 'test-provider' };
+  const captured = { agentConfig: null, chatArgs: null, registeredTools: [] };
+  const fw = {
+    getCwd: () => process.cwd(),
+    pluginManager: {
+      get: (name) => {
+        if (name === 'ai') return { getConfig: () => llmConfig };
+        return null;
+      },
+    },
+    // ★ Plugin base class 的 tool.register 会先 push 到 _registeredTools，
+    //   然后查 registry.has()，has() 返回 false 才调 registry.register()。
+    //   所以 mock 必须实现 has() 否则注册流程被中断。
+    registerTool: (toolDef) => {
+      captured.registeredTools.push(toolDef);
+    },
+    toolRegistry: {
+      has: () => false,
+      register: (toolDef) => {
+        captured.registeredTools.push(toolDef);
+      },
+    },
+    createSubAgent: (config) => {
+      captured.agentConfig = config;
+      return {
+        chat: async (msg, opts) => {
+          captured.chatArgs = { msg, opts };
+          return { success: true, message: 'mock reply', steps: 1 };
+        },
+      };
+    },
+  };
+  return { fw, captured };
+}
+
+function getChatTool(captured) {
+  return captured.registeredTools.find(t => t.name === 'chat');
+}
+
+describe('file-system 插件的 chat 工具', () => {
+  it('★ 关键：name 唯一（用 Date.now + random 避免冲突）', async () => {
+    const { fw, captured } = makeFakeFw();
+    const p = new FileSystemPlugin();
+    p.install(fw);
+    p.start(fw);
+    // 直接调工具的 execute
+    // find chat tool in _registeredTools
+    const chatTool = getChatTool(captured);
+    expect(chatTool).toBeTruthy();
+    if (!chatTool) return;
+
+    const r1 = await chatTool.execute({ message: 'hi' }, fw);
+    const r2 = await chatTool.execute({ message: 'hi2' }, fw);
+
+    expect(captured.agentConfig.name).toMatch(/^file-system-chat-\d+/);
+    // 两次调用应产生不同名字
+    const name1 = r1.metadata?.agent;
+    const name2 = r2.metadata?.agent;
+    expect(name1).not.toBe(name2);
+  });
+
+  it('★ 关键：sessionId 隔离（避免污染主对话）', async () => {
+    const { fw, captured } = makeFakeFw();
+    const p = new FileSystemPlugin();
+    p.install(fw);
+    p.start(fw);
+    const chatTool = getChatTool(captured);
+    if (!chatTool) throw new Error('no chat tool');
+
+    // 1. 不传 sessionId → 自动生成
+    await chatTool.execute({ message: 'm1' }, fw);
+    expect(captured.chatArgs.opts.sessionId).toMatch(/^chat-\d+$/);
+
+    // 2. 传 sessionId → 使用传入的
+    await chatTool.execute({ message: 'm2', sessionId: 'my-session' }, fw);
+    expect(captured.chatArgs.opts.sessionId).toBe('my-session');
+  });
+
+  it('★ 关键：跟随主 agent 的 LLM 配置', async () => {
+    const llmConfig = {
+      model: 'deepseek-chat',
+      provider: 'deepseek',
+      apiKey: 'sk-test',
+      baseURL: 'https://api.deepseek.com/v1',
+    };
+    const { fw, captured } = makeFakeFw({ llmConfig });
+    const p = new FileSystemPlugin();
+    p.install(fw);
+    p.start(fw);
+    const chatTool = getChatTool(captured);
+    if (!chatTool) throw new Error('no chat tool');
+
+    await chatTool.execute({ message: 'hi' }, fw);
+    expect(captured.agentConfig.model).toBe('deepseek-chat');
+    expect(captured.agentConfig.provider).toBe('deepseek');
+    expect(captured.agentConfig.apiKey).toBe('sk-test');
+    expect(captured.agentConfig.baseURL).toBe('https://api.deepseek.com/v1');
+    expect(captured.agentConfig.hidden).toBe(true);
+  });
+
+  it('★ 关键：data 只返回 message 文本（不返回整个对象）', async () => {
+    const { fw, captured } = makeFakeFw();
+    const p = new FileSystemPlugin();
+    p.install(fw);
+    p.start(fw);
+    const chatTool = getChatTool(captured);
+    if (!chatTool) throw new Error('no chat tool');
+
+    const r = await chatTool.execute({ message: 'hi' }, fw);
+    expect(r.success).toBe(true);
+    expect(r.data).toBe('mock reply');  // 只是字符串，不是整个对象
+    expect(typeof r.data).toBe('string');
+    // metadata 含元信息
+    expect(r.metadata).toBeTruthy();
+    expect(r.metadata.steps).toBe(1);
+  });
+
+  it('★ 关键：systemPrompt 默认值合理', async () => {
+    const { fw, captured } = makeFakeFw();
+    const p = new FileSystemPlugin();
+    p.install(fw);
+    p.start(fw);
+    const chatTool = getChatTool(captured);
+    if (!chatTool) throw new Error('no chat tool');
+
+    // 不传 systemPrompt → 用默认
+    await chatTool.execute({ message: 'hi' }, fw);
+    expect(captured.agentConfig.systemPrompt).toContain('文件系统');
+
+    // 传 systemPrompt → 覆盖
+    await chatTool.execute({
+      message: 'hi',
+      systemPrompt: '你是一个翻译助手',
+    }, fw);
+    expect(captured.agentConfig.systemPrompt).toBe('你是一个翻译助手');
+  });
+
+  it('★ 关键：错误捕获（AI 抛错时不崩溃）', async () => {
+    const registeredTools = [];
+    const fw = {
+      getCwd: () => process.cwd(),
+      pluginManager: { get: () => null },
+      registerTool: (t) => registeredTools.push(t),
+      toolRegistry: { has: () => false, register: (t) => registeredTools.push(t) },
+      createSubAgent: () => ({
+        chat: async () => { throw new Error('AI 服务不可用'); },
+      }),
+    };
+    const p = new FileSystemPlugin();
+    p.install(fw);
+    p.start(fw);
+    const chatTool = registeredTools.find(t => t.name === 'chat');
+    if (!chatTool) throw new Error('no chat tool');
+
+    const r = await chatTool.execute({ message: 'hi' }, fw);
+    expect(r.success).toBe(false);
+    expect(r.error).toBe('AI 服务不可用');
+  });
+
+  it('★ inputSchema 含 sessionId 字段', () => {
+    // 直接读源码验证
+    const src = fs.readFileSync(
+      path.join(__dirname, '../../plugins/io/file-system/index.js'),
+      'utf-8'
+    );
+    expect(src).toMatch(/sessionId:\s*z\.string\(\)\.optional\(\)/);
+  });
+
+  it('★ 删除死代码：const ai = framework.getAI()', () => {
+    const src = fs.readFileSync(
+      path.join(__dirname, '../../plugins/io/file-system/index.js'),
+      'utf-8'
+    );
+    expect(src).not.toMatch(/const\s+ai\s*=\s*framework\.getAI/);
+  });
+});