autosnippet 3.0.10 → 3.0.13

package/lib/service/chat/tools/guard.js

@@ -0,0 +1,265 @@
+/**
+ * guard.js — Guard 安全类工具 (6)
+ *
+ * 7b. list_guard_rules    列出 Guard 规则
+ * 8b. get_recommendations 获取推荐 Recipe
+ * 12. ai_translate        AI 翻译
+ * 13. guard_check_code    Guard 检查代码
+ * 14. query_violations    查询违规历史
+ * 15. generate_guard_rule AI 生成 Guard 规则
+ */
+
+// ────────────────────────────────────────────────────────────
+// 7b. list_guard_rules
+// ────────────────────────────────────────────────────────────
+export const listGuardRules = {
+  name: 'list_guard_rules',
+  description: '列出所有 Guard 规则（boundary-constraint 类型的 Recipe）。支持按语言/状态过滤。',
+  parameters: {
+    type: 'object',
+    properties: {
+      language: { type: 'string', description: '按语言过滤 (swift/objc 等)' },
+      includeBuiltIn: { type: 'boolean', description: '是否包含内置规则，默认 true' },
+      limit: { type: 'number', description: '返回数量上限，默认 50' },
+    },
+  },
+  handler: async (params, ctx) => {
+    const { language, includeBuiltIn = true, limit = 50 } = params;
+    const results = [];
+
+    // 数据库自定义规则
+    try {
+      const guardService = ctx.container.get('guardService');
+      const dbRules = await guardService.listRules({}, { page: 1, pageSize: limit });
+      results.push(...(dbRules.data || dbRules.items || []));
+    } catch {
+      /* not available */
+    }
+
+    // 内置规则
+    if (includeBuiltIn) {
+      try {
+        const guardCheckEngine = ctx.container.get('guardCheckEngine');
+        const builtIn = guardCheckEngine
+          .getRules(language || null)
+          .filter((r) => r.source === 'built-in');
+        results.push(...builtIn);
+      } catch {
+        /* not available */
+      }
+    }
+
+    return { total: results.length, rules: results.slice(0, limit) };
+  },
+};
+
+// ────────────────────────────────────────────────────────────
+// 8b. get_recommendations
+// ────────────────────────────────────────────────────────────
+export const getRecommendations = {
+  name: 'get_recommendations',
+  description: '获取推荐的 Recipe 列表（基于使用频率和质量排序）。',
+  parameters: {
+    type: 'object',
+    properties: {
+      limit: { type: 'number', description: '返回数量，默认 10' },
+    },
+  },
+  handler: async (params, ctx) => {
+    const knowledgeService = ctx.container.get('knowledgeService');
+    // V3: 推荐 = 活跃条目按使用量排序
+    return knowledgeService.list(
+      { lifecycle: 'active' },
+      { page: 1, pageSize: params.limit || 10 }
+    );
+  },
+};
+
+// ────────────────────────────────────────────────────────────
+// 12. ai_translate
+// ────────────────────────────────────────────────────────────
+export const aiTranslate = {
+  name: 'ai_translate',
+  description: 'AI 翻译 — 将中文 summary/usageGuide 翻译为英文。',
+  parameters: {
+    type: 'object',
+    properties: {
+      summary: { type: 'string', description: '中文摘要' },
+      usageGuide: { type: 'string', description: '中文使用指南' },
+    },
+  },
+  handler: async (params, ctx) => {
+    if (!ctx.aiProvider) {
+      return { error: 'AI provider not available' };
+    }
+    const { summary, usageGuide } = params;
+    if (!summary && !usageGuide) {
+      return { summaryEn: '', usageGuideEn: '' };
+    }
+
+    const systemPrompt =
+      'You are a technical translator. Translate from Chinese to English. Keep technical terms unchanged. Return ONLY valid JSON: { "summaryEn": "...", "usageGuideEn": "..." }.';
+    const parts = [];
+    if (summary) {
+      parts.push(`summary: ${summary}`);
+    }
+    if (usageGuide) {
+      parts.push(`usageGuide: ${usageGuide}`);
+    }
+
+    const parsed = await ctx.aiProvider.chatWithStructuredOutput(parts.join('\n'), {
+      systemPrompt,
+      temperature: 0.2,
+    });
+    return parsed || { summaryEn: summary || '', usageGuideEn: usageGuide || '' };
+  },
+};
+
+// ────────────────────────────────────────────────────────────
+// 13. guard_check_code
+// ────────────────────────────────────────────────────────────
+export const guardCheckCode = {
+  name: 'guard_check_code',
+  description: '对代码运行 Guard 规则检查，返回违规列表（支持内置规则 + 数据库自定义规则）。',
+  parameters: {
+    type: 'object',
+    properties: {
+      code: { type: 'string', description: '待检查的源代码' },
+      language: { type: 'string', description: '编程语言 (swift/objc/javascript 等)' },
+      scope: { type: 'string', description: '检查范围 (file/target/project)，默认 file' },
+    },
+    required: ['code'],
+  },
+  handler: async (params, ctx) => {
+    const { code, language, scope = 'file' } = params;
+
+    // 优先用 GuardCheckEngine（内置 + DB 规则）
+    try {
+      const engine = ctx.container.get('guardCheckEngine');
+      const violations = engine.checkCode(code, language || 'unknown', { scope });
+      // reasoning 已由 GuardCheckEngine.checkCode() 内置附加
+      return { violationCount: violations.length, violations };
+    } catch {
+      /* not available */
+    }
+
+    // 降级到 GuardService.checkCode（仅 DB 规则）
+    try {
+      const guardService = ctx.container.get('guardService');
+      const matches = await guardService.checkCode(code, { language });
+      return { violationCount: matches.length, violations: matches };
+    } catch (err) {
+      return { error: err.message };
+    }
+  },
+};
+
+// ────────────────────────────────────────────────────────────
+// 14. query_violations
+// ────────────────────────────────────────────────────────────
+export const queryViolations = {
+  name: 'query_violations',
+  description: '查询 Guard 违规历史记录和统计。',
+  parameters: {
+    type: 'object',
+    properties: {
+      file: { type: 'string', description: '按文件路径过滤' },
+      limit: { type: 'number', description: '返回数量，默认 20' },
+      statsOnly: { type: 'boolean', description: '仅返回统计数据，默认 false' },
+    },
+  },
+  handler: async (params, ctx) => {
+    const { file, limit = 20, statsOnly = false } = params;
+    const store = ctx.container.get('violationsStore');
+
+    if (statsOnly) {
+      return store.getStats();
+    }
+
+    if (file) {
+      return { runs: store.getRunsByFile(file) };
+    }
+
+    return store.list({}, { page: 1, limit });
+  },
+};
+
+// ────────────────────────────────────────────────────────────
+// 15. generate_guard_rule
+// ────────────────────────────────────────────────────────────
+export const generateGuardRule = {
+  name: 'generate_guard_rule',
+  description: 'AI 生成 Guard 规则 — 描述你想阻止的代码模式，AI 自动生成正则表达式和规则定义。',
+  parameters: {
+    type: 'object',
+    properties: {
+      description: {
+        type: 'string',
+        description: '规则描述（例如 "禁止在主线程使用同步网络请求"）',
+      },
+      language: { type: 'string', description: '目标语言 (swift/objc 等)' },
+      severity: { type: 'string', description: '严重程度 (error/warning/info)，默认 warning' },
+      autoCreate: { type: 'boolean', description: '是否自动创建到数据库，默认 false' },
+    },
+    required: ['description'],
+  },
+  handler: async (params, ctx) => {
+    if (!ctx.aiProvider) {
+      return { error: 'AI provider not available' };
+    }
+    const { description, language = 'swift', severity = 'warning', autoCreate = false } = params;
+
+    const prompt = `Generate a Guard rule for this requirement:
+Description: ${description}
+Language: ${language}
+Severity: ${severity}
+
+Return ONLY valid JSON:
+{
+  "name": "rule-name-kebab-case",
+  "description": "One-line description in English",
+  "description_cn": "一行中文描述",
+  "pattern": "regex pattern for matching the problematic code",
+  "languages": ["${language}"],
+  "severity": "${severity}",
+  "testCases": {
+    "shouldMatch": ["code example that should trigger"],
+    "shouldNotMatch": ["code example that should NOT trigger"]
+  }
+}`;
+
+    const rule = await ctx.aiProvider.chatWithStructuredOutput(prompt, { temperature: 0.2 });
+    if (!rule) {
+      return { error: 'Failed to parse AI response' };
+    }
+
+    // 验证正则表达式
+    try {
+      new RegExp(rule.pattern);
+    } catch (e) {
+      return { error: `Invalid regex pattern: ${e.message}`, rule };
+    }
+
+    // 自动创建
+    if (autoCreate && rule.name && rule.pattern) {
+      try {
+        const guardService = ctx.container.get('guardService');
+        const created = await guardService.createRule(
+          {
+            name: rule.name,
+            description: rule.description || description,
+            pattern: rule.pattern,
+            languages: rule.languages || [language],
+            severity: rule.severity || severity,
+          },
+          { userId: 'agent' }
+        );
+        return { rule, created: true, recipeId: created.id };
+      } catch (err) {
+        return { rule, created: false, error: err.message };
+      }
+    }
+
+    return { rule, created: false };
+  },
+};

package/lib/service/chat/tools/index.js

@@ -0,0 +1,250 @@
+/**
+ * tools/index.js — Barrel 导出文件
+ *
+ * 从各子模块导入所有工具，按原始顺序组装 ALL_TOOLS 数组并导出。
+ */
+
+// ── 项目数据访问 (5) ──
+import {
+  searchProjectCode,
+  readProjectFile,
+  listProjectStructure,
+  getFileSummary,
+  semanticSearchCode,
+} from './project-access.js';
+
+// ── 查询类 (6) ──
+import {
+  searchRecipes,
+  searchCandidates,
+  getRecipeDetail,
+  getProjectStats,
+  searchKnowledge,
+  getRelatedRecipes,
+} from './query.js';
+
+// ── AI 分析类 (4) ──
+import {
+  summarizeCode,
+  extractRecipes,
+  enrichCandidate,
+  refineBootstrapCandidates,
+} from './ai-analysis.js';
+
+// ── Guard 安全类 (6) ──
+import {
+  listGuardRules,
+  getRecommendations,
+  aiTranslate,
+  guardCheckCode,
+  queryViolations,
+  generateGuardRule,
+} from './guard.js';
+
+// ── 知识图谱类 (3) ──
+import {
+  checkDuplicate,
+  discoverRelations,
+  addGraphEdge,
+} from './knowledge-graph.js';
+
+// ── 生命周期操作类 (11) ──
+import {
+  submitCandidate,
+  saveDocument,
+  approveCandidate,
+  rejectCandidate,
+  publishRecipe,
+  deprecateRecipe,
+  updateRecipe,
+  recordUsage,
+  qualityScore,
+  validateCandidate,
+  getFeedbackStats,
+} from './lifecycle.js';
+
+// ── 基础设施类 (7) ──
+import {
+  graphImpactAnalysis,
+  rebuildIndex,
+  queryAuditLog,
+  loadSkill,
+  createSkillTool,
+  suggestSkills,
+  bootstrapKnowledgeTool,
+} from './infrastructure.js';
+
+// ── 组合工具 + 元工具 (6) ──
+import {
+  analyzeCode,
+  knowledgeOverview,
+  submitWithCheck,
+  getToolDetails,
+  planTask,
+  reviewMyOutput,
+} from './composite.js';
+
+// ── AST 结构化分析 + Agent Memory (10) ──
+import {
+  getProjectOverview,
+  getClassHierarchy,
+  getClassInfo,
+  getProtocolInfo,
+  getMethodOverrides,
+  getCategoryMap,
+  getPreviousAnalysis,
+  noteFinding,
+  getPreviousEvidence,
+  queryCodeGraph,
+} from './ast-graph.js';
+
+// ── Re-export 所有工具 ──
+export {
+  // 项目数据访问
+  searchProjectCode,
+  readProjectFile,
+  listProjectStructure,
+  getFileSummary,
+  semanticSearchCode,
+  // 查询类
+  searchRecipes,
+  searchCandidates,
+  getRecipeDetail,
+  getProjectStats,
+  searchKnowledge,
+  getRelatedRecipes,
+  // AI 分析类
+  summarizeCode,
+  extractRecipes,
+  enrichCandidate,
+  refineBootstrapCandidates,
+  // Guard 安全类
+  listGuardRules,
+  getRecommendations,
+  aiTranslate,
+  guardCheckCode,
+  queryViolations,
+  generateGuardRule,
+  // 知识图谱类
+  checkDuplicate,
+  discoverRelations,
+  addGraphEdge,
+  // 生命周期操作类
+  submitCandidate,
+  saveDocument,
+  approveCandidate,
+  rejectCandidate,
+  publishRecipe,
+  deprecateRecipe,
+  updateRecipe,
+  recordUsage,
+  qualityScore,
+  validateCandidate,
+  getFeedbackStats,
+  // 基础设施类
+  graphImpactAnalysis,
+  rebuildIndex,
+  queryAuditLog,
+  loadSkill,
+  createSkillTool,
+  suggestSkills,
+  bootstrapKnowledgeTool,
+  // 组合工具 + 元工具
+  analyzeCode,
+  knowledgeOverview,
+  submitWithCheck,
+  getToolDetails,
+  planTask,
+  reviewMyOutput,
+  // AST 结构化分析
+  getProjectOverview,
+  getClassHierarchy,
+  getClassInfo,
+  getProtocolInfo,
+  getMethodOverrides,
+  getCategoryMap,
+  getPreviousAnalysis,
+  // Agent Memory
+  noteFinding,
+  getPreviousEvidence,
+  // 代码实体图谱
+  queryCodeGraph,
+};
+
+// ── ALL_TOOLS 数组（与原始 tools.js 顺序一致）──
+export const ALL_TOOLS = [
+  // 项目数据访问 (5) — 含 v10 Agent-Pull 工具
+  searchProjectCode,
+  readProjectFile,
+  listProjectStructure,
+  getFileSummary,
+  semanticSearchCode,
+  // 查询类 (8)
+  searchRecipes,
+  searchCandidates,
+  getRecipeDetail,
+  getProjectStats,
+  searchKnowledge,
+  getRelatedRecipes,
+  listGuardRules,
+  getRecommendations,
+  // AI 分析类 (5)
+  summarizeCode,
+  extractRecipes,
+  enrichCandidate,
+  refineBootstrapCandidates,
+  aiTranslate,
+  // Guard 安全类 (3)
+  guardCheckCode,
+  queryViolations,
+  generateGuardRule,
+  // 生命周期操作类 (7)
+  submitCandidate,
+  saveDocument,
+  approveCandidate,
+  rejectCandidate,
+  publishRecipe,
+  deprecateRecipe,
+  updateRecipe,
+  recordUsage,
+  // 质量与反馈类 (3)
+  qualityScore,
+  validateCandidate,
+  getFeedbackStats,
+  // 知识图谱类 (3)
+  checkDuplicate,
+  discoverRelations,
+  addGraphEdge,
+  // 基础设施类 (3)
+  graphImpactAnalysis,
+  rebuildIndex,
+  queryAuditLog,
+  // Skills & Bootstrap (4)
+  loadSkill,
+  createSkillTool,
+  suggestSkills,
+  bootstrapKnowledgeTool,
+  // 组合工具 (3) — 减少 ReAct 轮次
+  analyzeCode,
+  knowledgeOverview,
+  submitWithCheck,
+  // 元工具 (3) — Agent 自主能力增强
+  getToolDetails,
+  planTask,
+  reviewMyOutput,
+  // AST 结构化分析 (7) — v3.0 AI-First Bootstrap
+  getProjectOverview,
+  getClassHierarchy,
+  getClassInfo,
+  getProtocolInfo,
+  getMethodOverrides,
+  getCategoryMap,
+  getPreviousAnalysis,
+  // Agent Memory 增强 (2) — 工作记忆 + 情景记忆
+  noteFinding,
+  getPreviousEvidence,
+  // 代码实体图谱 (1) — Phase E
+  queryCodeGraph,
+];
+
+export default ALL_TOOLS;