preflight-mcp 0.1.9 → 0.2.1

package/README.md

@@ -16,6 +16,10 @@ Each bundle contains:
 ## Features
 
 - **13 MCP tools** to create/update/repair/search/read bundles, generate evidence graphs, and manage trace links
+- **5 MCP prompts** for interactive guidance: menu, analyze guide, search guide, manage guide, trace guide
+- **LLM-friendly outputs**: After bundle creation, prompts user to generate dependency graph for deeper analysis
+- **Proactive trace links**: LLM automatically discovers and records code↔test, code↔doc relationships
+- **Auto-export trace.json**: Trace links are automatically exported to JSON for direct LLM reading (no API needed)
 - **Progress tracking**: Real-time progress reporting for long-running operations (create/update bundles)
 - **Bundle integrity check**: Prevents operations on incomplete bundles with helpful error messages
 - **De-duplication with in-progress lock**: Prevent duplicate bundle creation even during MCP timeouts
@@ -151,10 +155,11 @@ Input (example):
 
 ### `preflight_read_file`
 Read file(s) from bundle. Two modes:
-- **Batch mode** (omit `file`): Returns ALL key files (OVERVIEW.md, START_HERE.md, AGENTS.md, manifest.json, repo READMEs) in one call
-- **Single file mode** (provide `file`): Returns that specific file
-- Triggers: "查看bundle", "bundle概览", "项目信息", "show bundle"
+- **Batch mode** (omit `file`): Returns ALL key files (OVERVIEW.md, START_HERE.md, AGENTS.md, manifest.json, deps/dependency-graph.json, repo READMEs) in one call
+- **Single file mode** (provide `file`): Returns that specific file (e.g., `deps/dependency-graph.json` for dependency graph)
+- Triggers: "查看bundle", "bundle概览", "项目信息", "show bundle", "读取依赖图"
 - Use `file: "manifest.json"` to get bundle metadata (repos, timestamps, tags, etc.)
+- Use `file: "deps/dependency-graph.json"` to read the dependency graph (generated by `preflight_evidence_dependency_graph`)
 
 ### `preflight_delete_bundle`
 Delete/remove a bundle permanently.
@@ -203,10 +208,16 @@ Generate an evidence-based dependency graph. Two modes:
 - Emits `imports` edges (file → module) and `imports_resolved` edges (file → internal file).
 
 ### `preflight_trace_upsert`
-Upsert traceability links (commit↔ticket, symbol↔test, code↔doc, etc.) for a bundle.
+Create or update traceability links (code↔test, code↔doc, file↔requirement).
+- **Proactive use**: LLM automatically records discovered relationships during code analysis
+- Common link types: `tested_by`, `implements`, `documents`, `relates_to`, `depends_on`
+- **Auto-exports** to `trace/trace.json` after each upsert for direct LLM reading
 
 ### `preflight_trace_query`
-Query traceability links (fast when `bundleId` is provided; can scan across bundles when omitted).
+Query traceability links (code↔test, code↔doc, commit↔ticket).
+- **Proactive use**: LLM automatically queries trace links when analyzing specific files
+- Helps answer: "Does this code have tests?", "What requirements does this implement?"
+- Fast when `bundleId` is provided; can scan across bundles when omitted.
 
 ### `preflight_cleanup_orphans`
 Remove incomplete or corrupted bundles (bundles without valid manifest.json).
@@ -224,6 +235,33 @@ Check status of bundle creation/update tasks (progress tracking).
 - Query by `taskId` (from error), `fingerprint`, or `repos`
 - Shows: phase, progress percentage, message, elapsed time
 
+## Prompts (5 total)
+
+MCP prompts provide interactive guidance for users. Call these to get usage instructions and example prompts.
+
+### `preflight_menu`
+Main menu showing all Preflight features.
+- Triggers: "preflight有什么功能", "有什么工具", "what can preflight do", "show menu"
+
+### `preflight_analyze_guide`
+Deep analysis guide with step-by-step workflow and copyable prompts.
+- Shows: Bundle file structure, recommended analysis flow, example prompts
+- Args: `projectPath` (optional)
+
+### `preflight_search_guide`
+Search functionality guide.
+- Shows: Single bundle search, cross-bundle search by tags, FTS5 syntax tips
+- Args: `bundleId` (optional)
+
+### `preflight_manage_guide`
+Bundle management operations guide.
+- Shows: List, view, update, repair, delete bundle operations
+
+### `preflight_trace_guide`
+Traceability links guide.
+- Shows: Query and create code↔test, code↔doc relationships
+- Args: `bundleId` (optional)
+
 ## Resources
 
 ### `preflight://bundles`
@@ -287,7 +325,9 @@ Inside a bundle directory:
 - `OVERVIEW.md`
 - `indexes/search.sqlite3`
 - `analysis/FACTS.json` (static analysis)
+- `deps/dependency-graph.json` (global import graph; generated on demand)
 - `trace/trace.sqlite3` (traceability links; created on demand)
+- `trace/trace.json` (**NEW**: auto-exported JSON for direct LLM reading)
 - `repos/<owner>/<repo>/raw/...`
 - `repos/<owner>/<repo>/norm/...` (GitHub/local snapshots)
 - `libraries/context7/<...>/meta.json`

package/README.zh-CN.md

@@ -15,7 +15,8 @@
 
 ## Features
 
-- **12 个 MCP 工具**：create/update/repair/search/evidence/trace/read/cleanup（外加 resources）
+- **13 个 MCP 工具**：create/update/repair/search/evidence/trace/read/cleanup（外加 resources）
+- **5 个 MCP prompts**：交互式引导（菜单、分析指南、搜索指南、管理指南、追溯指南）
 - **去重**：避免对相同的规范化输入重复索引
 - **可靠的 GitHub 获取**：可配置 git clone 超时 + GitHub archive（zipball）兜底
 - **离线修复**：无需重新抓取，重建缺失/为空的派生物（index/guides/overview）
@@ -168,9 +169,12 @@ npm run smoke
 **💡 提示**：对于代码仓库，创建 bundle 后可进一步使用 `preflight_evidence_dependency_graph` 获取依赖图，或使用 `preflight_trace_upsert` 记录代码←→需求/测试的追溯链接。
 
 ### `preflight_read_file`
-从 bundle 读取文件（OVERVIEW.md、START_HERE.md、AGENTS.md、manifest.json 或任何仓库文件）。
-- 触发词：「查看概览」「项目概览」「看README」「bundle详情」「bundle状态」「仓库信息」
-- **注意**：使用 `file="manifest.json"` 可获取完整的 bundle 元信息（替代原 `preflight_bundle_info`）
+从 bundle 读取文件。两种模式：
+- **批量模式**（省略 `file`）：返回所有关键文件（OVERVIEW.md、START_HERE.md、AGENTS.md、manifest.json、deps/dependency-graph.json、repo READMEs）
+- **单文件模式**（提供 `file`）：返回指定文件（如 `deps/dependency-graph.json` 获取依赖图）
+- 触发词：「查看概览」「项目概览」「bundle详情」「读取依赖图」
+- 使用 `file: "manifest.json"` 获取 bundle 元数据
+- 使用 `file: "deps/dependency-graph.json"` 读取依赖图（由 `preflight_evidence_dependency_graph` 生成）
 
 ### `preflight_delete_bundle`
 永久删除/移除一个 bundle。
@@ -232,6 +236,39 @@ npm run smoke
 - 输出：`totalFound`, `totalCleaned`, `details`
 - 说明：服务启动时也会自动执行后台清理（非阻塞）
 
+### `preflight_get_task_status`
+检查 bundle 创建/更新任务的状态（进度追踪）。
+- 触发词：「查看进度」「任务状态」「下载进度」
+- 通过 `taskId`、`fingerprint` 或 `repos` 查询
+- 显示：阶段、进度百分比、消息、已用时间
+
+## Prompts（5 个）
+
+MCP prompts 提供交互式引导。调用这些 prompt 获取使用说明和示例。
+
+### `preflight_menu`
+主菜单，显示所有 Preflight 功能。
+- 触发词：「preflight有什么功能」「有什么工具」「what can preflight do」
+
+### `preflight_analyze_guide`
+深入分析指南，包含分步流程和可复制的 prompt。
+- 显示：Bundle 文件结构、推荐分析流程、示例 prompt
+- 参数：`projectPath`（可选）
+
+### `preflight_search_guide`
+搜索功能指南。
+- 显示：单 bundle 搜索、跨 bundle 按标签搜索、FTS5 语法提示
+- 参数：`bundleId`（可选）
+
+### `preflight_manage_guide`
+Bundle 管理操作指南。
+- 显示：列出、查看、更新、修复、删除 bundle 操作
+
+### `preflight_trace_guide`
+追溯链接指南。
+- 显示：查询和创建代码↔测试、代码↔文档关系
+- 参数：`bundleId`（可选）
+
 ## Resources
 
 ### `preflight://bundles`
@@ -296,7 +333,9 @@ bundle 目录内部：
 - `OVERVIEW.md`
 - `indexes/search.sqlite3`
 - **`analysis/FACTS.json`**（静态分析）
+- **`deps/dependency-graph.json`**（全局依赖图；按需生成）
 - `trace/trace.sqlite3`（traceability links；按需创建）
+- `trace/trace.json`（自动导出的 JSON，便于 LLM 直接读取）
 - `repos/<owner>/<repo>/raw/...`
 - `repos/<owner>/<repo>/norm/...`（GitHub/local 快照）
 - `libraries/context7/<...>/meta.json`

package/dist/server.js

@@ -100,7 +100,7 @@ export async function startServer() {
     startHttpServer(cfg);
     const server = new McpServer({
         name: 'preflight-mcp',
-        version: '0.1.8',
+        version: '0.2.1',
         description: 'Create evidence-based preflight bundles for repositories (docs + code) with SQLite FTS search.',
     }, {
         capabilities: {
@@ -108,6 +108,10 @@ export async function startServer() {
                 // We can emit list changed notifications when new bundles appear.
                 listChanged: true,
             },
+            prompts: {
+                // We provide interactive guidance prompts.
+                listChanged: true,
+            },
         },
     });
     // Resource template to read any file inside a bundle.
@@ -261,12 +265,12 @@ export async function startServer() {
     server.registerTool('preflight_read_file', {
         title: 'Read bundle file(s)',
         description: 'Read file(s) from bundle. Two modes: ' +
-            '(1) Omit "file" param → returns ALL key files (OVERVIEW.md, START_HERE.md, AGENTS.md, manifest.json, repo READMEs) in one call. ' +
-            '(2) Provide "file" param → returns that specific file. ' +
-            'Use when: "查看bundle", "show bundle", "read overview", "bundle概览", "项目信息".',
+            '(1) Omit "file" param → returns ALL key files (OVERVIEW.md, START_HERE.md, AGENTS.md, manifest.json, deps/dependency-graph.json, repo READMEs) in one call. ' +
+            '(2) Provide "file" param → returns that specific file (e.g., "deps/dependency-graph.json" for the dependency graph). ' +
+            'Use when: "查看bundle", "show bundle", "read overview", "bundle概览", "项目信息", "读取依赖图", "show dependency graph".',
         inputSchema: {
             bundleId: z.string().describe('Bundle ID to read.'),
-            file: z.string().optional().describe('Specific file to read. If omitted, returns all key files (OVERVIEW.md, START_HERE.md, AGENTS.md, manifest.json, repo READMEs).'),
+            file: z.string().optional().describe('Specific file to read (e.g., "deps/dependency-graph.json"). If omitted, returns all key files including dependency graph if exists.'),
         },
         outputSchema: {
             bundleId: z.string(),
@@ -296,7 +300,7 @@ export async function startServer() {
                 };
             }
             // Batch mode: read all key files
-            const keyFiles = ['OVERVIEW.md', 'START_HERE.md', 'AGENTS.md', 'manifest.json'];
+            const keyFiles = ['OVERVIEW.md', 'START_HERE.md', 'AGENTS.md', 'manifest.json', 'deps/dependency-graph.json'];
             const files = {};
             for (const file of keyFiles) {
                 try {
@@ -457,7 +461,12 @@ export async function startServer() {
                 textResponse += '\n';
             }
             textResponse += `✅ Bundle created: ${summary.bundleId}\n`;
-            textResponse += `Repos: ${summary.repos.map(r => `${r.id} (${r.source})`).join(', ')}`;
+            textResponse += `Repos: ${summary.repos.map(r => `${r.id} (${r.source})`).join(', ')}\n\n`;
+            // Prompt user for advanced analysis
+            textResponse += `📊 **Recommended next steps:**\n`;
+            textResponse += `Would you like me to generate a **global dependency graph** for deeper code analysis? ` +
+                `This will analyze import relationships across all files.\n`;
+            textResponse += `(Call \`preflight_evidence_dependency_graph\` with this bundleId to generate)`;
             return {
                 content: [{ type: 'text', text: textResponse }],
                 structuredContent: out,
@@ -792,8 +801,9 @@ export async function startServer() {
         }
     });
     server.registerTool('preflight_evidence_dependency_graph', {
-        title: 'Evidence: dependency graph (callers + imports)',
-        description: 'Generate an evidence-based dependency graph. IMPORTANT: Before running, ASK the user which bundle and which file/mode they want! ' +
+        title: 'Evidence: dependency graph',
+        description: '**Proactive use recommended**: Generate dependency graphs to understand code structure. ' +
+            'Generate an evidence-based dependency graph. IMPORTANT: Before running, ASK the user which bundle and which file/mode they want! ' +
             'Two modes: (1) TARGET MODE: analyze a specific file (provide target.file). (2) GLOBAL MODE: project-wide graph (omit target). ' +
             'Do NOT automatically choose bundle or mode - confirm with user first! ' +
             'File path must be bundle-relative: repos/{owner}/{repo}/norm/{path}.',
@@ -822,7 +832,12 @@ export async function startServer() {
     });
     server.registerTool('preflight_trace_upsert', {
         title: 'Trace: upsert links',
-        description: 'Upsert traceability links (commit↔ticket, symbol↔test, code↔doc, etc.) for a bundle. Stores trace edges in a per-bundle SQLite database.',
+        description: 'Create or update traceability links (code↔test, code↔doc, file↔requirement). ' +
+            '**Proactive use recommended**: When you discover relationships during code analysis ' +
+            '(e.g., "this file has a corresponding test", "this module implements feature X"), ' +
+            'automatically create trace links to record these findings for future queries. ' +
+            'Common link types: tested_by, implements, documents, relates_to, depends_on. ' +
+            'Stores trace edges in a per-bundle SQLite database.',
         inputSchema: TraceUpsertInputSchema,
         outputSchema: {
             bundleId: z.string(),
@@ -848,7 +863,11 @@ export async function startServer() {
     });
     server.registerTool('preflight_trace_query', {
         title: 'Trace: query links',
-        description: 'Query traceability links. Provide bundleId for fast queries; if omitted, scans across bundles (capped). This tool is read-only.',
+        description: 'Query traceability links (code↔test, code↔doc, commit↔ticket). ' +
+            '**Proactive use recommended**: When analyzing a specific file or discussing code structure, ' +
+            'automatically query trace links to find related tests, documentation, or requirements. ' +
+            'This helps answer questions like "does this code have tests?" or "what requirements does this implement?". ' +
+            'Provide bundleId for fast queries; if omitted, scans across bundles (capped). This tool is read-only.',
         inputSchema: TraceQueryInputSchema,
         outputSchema: {
             bundleId: z.string().optional(),
@@ -1033,6 +1052,241 @@ export async function startServer() {
             throw wrapPreflightError(err);
         }
     });
+    // ============================================================
+    // PROMPTS - Interactive guidance for users
+    // ============================================================
+    // Main menu prompt - shows all available features
+    server.registerPrompt('preflight_menu', {
+        title: 'Preflight 功能菜单',
+        description: '显示 Preflight 所有可用功能的交互式菜单。Use when: "preflight有什么功能", "有什么工具", "what can preflight do", "show menu".',
+    }, async () => {
+        return {
+            messages: [
+                {
+                    role: 'user',
+                    content: {
+                        type: 'text',
+                        text: `🛠️ **Preflight 功能菜单**
+
+请选择您需要的功能：
+
+**1. 📂 深入分析项目**
+创建 bundle 并生成全局依赖图，理解代码架构
+
+**2. 🔍 搜索代码/文档**
+在已索引的项目中全文搜索代码和文档
+
+**3. 📋 管理 bundles**
+列出、更新、修复、删除已有的 bundle
+
+**4. 🔗 追溯链接**
+查询/创建代码-测试-文档之间的关联关系
+
+---
+请输入功能编号 (1-4) 或直接描述您的需求。`,
+                    },
+                },
+            ],
+        };
+    });
+    // Deep analysis guide prompt
+    server.registerPrompt('preflight_analyze_guide', {
+        title: '深入分析项目指南',
+        description: '提供深入分析项目的操作指南和示例 prompt。Use when user selected "深入分析" or wants to analyze a project.',
+        argsSchema: {
+            projectPath: z.string().optional().describe('项目路径或 GitHub 仓库地址（可选）'),
+        },
+    }, async (args) => {
+        const pathExample = args.projectPath || 'E:\\coding\\my-project 或 owner/repo';
+        return {
+            messages: [
+                {
+                    role: 'user',
+                    content: {
+                        type: 'text',
+                        text: `📊 **深入分析项目指南**
+
+**第一步：提供项目路径**
+- 本地路径：\`E:\\coding\\my-project\`
+- GitHub：\`owner/repo\` 或完整 URL
+
+**第二步：复制以下完美 prompt（送给工作 LLM）**
+
+\`\`\`
+请执行以下分析流程：
+
+1. 使用 preflight_create_bundle 创建 ${pathExample} 的 bundle
+2. 使用 preflight_evidence_dependency_graph 生成全局依赖图
+3. 使用 preflight_read_file 读取 bundle 内容，分析：
+   - OVERVIEW.md 了解项目概览
+   - deps/dependency-graph.json 查看依赖关系
+   - START_HERE.md 了解入口点
+
+然后总结：
+1. 项目核心功能是什么
+2. 主要模块及其关系（基于依赖图）
+3. 代码架构特点
+\`\`\`
+
+**Bundle 文件结构说明：**
+| 文件 | 内容 |
+|------|------|
+| \`OVERVIEW.md\` | 项目概览和结构总结 |
+| \`START_HERE.md\` | 入口文件和关键路径 |
+| \`AGENTS.md\` | AI Agent 使用指南 |
+| \`deps/dependency-graph.json\` | 依赖关系图（节点+边） |
+| \`manifest.json\` | bundle 元数据 |
+| \`repos/{owner}/{repo}/norm/\` | 规范化源代码 |
+
+---
+💡 提示：搜索功能只能查找代码文件内容，不能搜索依赖图。要查看依赖图，请用 preflight_read_file 读取 deps/dependency-graph.json。`,
+                    },
+                },
+            ],
+        };
+    });
+    // Search guide prompt
+    server.registerPrompt('preflight_search_guide', {
+        title: '搜索代码/文档指南',
+        description: '提供搜索功能的操作指南和示例 prompt。Use when user selected "搜索" or wants to search in bundles.',
+        argsSchema: {
+            bundleId: z.string().optional().describe('要搜索的 bundle ID（可选）'),
+        },
+    }, async (args) => {
+        const bundleHint = args.bundleId ? `bundle \`${args.bundleId}\`` : '指定的 bundle';
+        return {
+            messages: [
+                {
+                    role: 'user',
+                    content: {
+                        type: 'text',
+                        text: `🔍 **搜索代码/文档指南**
+
+**搜索模式：**
+
+**1. 单 bundle 搜索**（需要先知道 bundleId）
+\`\`\`
+在 ${bundleHint} 中搜索 "config parser"
+\`\`\`
+
+**2. 跨 bundle 搜索**（按 tags 过滤）
+\`\`\`
+在所有 MCP 相关项目中搜索 "tool registration"
+在标签为 agent 的项目中搜索 "LLM"
+\`\`\`
+
+**3. 列出所有 bundle**（不确定有哪些时）
+\`\`\`
+列出所有 bundle
+或: preflight list bundles
+\`\`\`
+
+---
+💡 搜索支持 FTS5 全文语法，如：\`config AND parser\`、\`"exact phrase"\``,
+                    },
+                },
+            ],
+        };
+    });
+    // Manage bundles guide prompt
+    server.registerPrompt('preflight_manage_guide', {
+        title: '管理 bundles 指南',
+        description: '提供 bundle 管理操作的指南。Use when user selected "管理" or wants to manage bundles.',
+    }, async () => {
+        return {
+            messages: [
+                {
+                    role: 'user',
+                    content: {
+                        type: 'text',
+                        text: `📋 **管理 Bundles 指南**
+
+**常用操作：**
+
+**列出所有 bundle**
+\`\`\`
+列出所有 bundle
+或: 查看有哪些项目已索引
+\`\`\`
+
+**查看 bundle 详情**
+\`\`\`
+查看 bundle {bundleId} 的概览
+或: 读取 bundle {bundleId}
+\`\`\`
+
+**更新 bundle**（同步最新代码）
+\`\`\`
+更新 bundle {bundleId}
+或: 检查 {bundleId} 是否有更新
+\`\`\`
+
+**修复 bundle**（重建索引）
+\`\`\`
+修复 bundle {bundleId}
+或: 重建 {bundleId} 的搜索索引
+\`\`\`
+
+**删除 bundle**
+\`\`\`
+删除 bundle {bundleId}
+\`\`\`
+
+---
+💡 先运行「列出所有 bundle」获取 bundleId 列表`,
+                    },
+                },
+            ],
+        };
+    });
+    // Trace guide prompt
+    server.registerPrompt('preflight_trace_guide', {
+        title: '追溯链接指南',
+        description: '提供代码追溯功能的操作指南。Use when user selected "追溯" or wants to trace code relationships.',
+        argsSchema: {
+            bundleId: z.string().optional().describe('bundle ID（可选）'),
+        },
+    }, async (args) => {
+        const bundleHint = args.bundleId || '{bundleId}';
+        return {
+            messages: [
+                {
+                    role: 'user',
+                    content: {
+                        type: 'text',
+                        text: `🔗 **追溯链接指南**
+
+追溯功能用于建立和查询代码之间的关联关系：
+- 代码 ↔ 测试
+- 代码 ↔ 文档
+- 模块 ↔ 需求
+
+**查询已有的追溯链接**
+\`\`\`
+查询 bundle ${bundleHint} 中 src/main.ts 的相关测试
+查询所有 implements 类型的追溯链接
+\`\`\`
+
+**创建追溯链接**
+\`\`\`
+在 bundle ${bundleHint} 中创建追溯：
+src/parser.ts 被 tests/parser.test.ts 测试
+\`\`\`
+
+**常用链接类型：**
+- \`tested_by\` - 被...测试
+- \`implements\` - 实现了...
+- \`documents\` - 文档描述了...
+- \`depends_on\` - 依赖于...
+- \`relates_to\` - 相关联
+
+---
+💡 追溯链接会持久化存储，便于未来快速查询代码关系`,
+                    },
+                },
+            ],
+        };
+    });
     // Provide backward-compatible parsing of the same URI via resources/read for clients that bypass templates.
     // This is a safety net: if a client gives us a fully-specified URI, we can still serve it.
     server.registerResource('bundle-file-compat', 'preflight://bundle-file', {

package/dist/trace/store.js

@@ -98,8 +98,56 @@ export async function upsertTraceEdges(traceDbPath, edges) {
             return ids;
         });
         const ids = tx(edges);
+        // Close DB before export (export opens its own readonly connection)
+        db.close();
+        // Auto-export to JSON after each upsert for LLM direct reading
+        try {
+            await exportTraceToJson(traceDbPath);
+        }
+        catch {
+            // Non-critical: JSON export failure shouldn't block upsert
+        }
         return { upserted: ids.length, ids };
     }
+    catch (err) {
+        db.close();
+        throw err;
+    }
+}
+/**
+ * Export all trace edges to a JSON file for LLM direct reading.
+ * Called automatically after upsertTraceEdges.
+ */
+export async function exportTraceToJson(traceDbPath) {
+    const jsonPath = traceDbPath.replace(/\.sqlite3$/, '.json');
+    const db = new Database(traceDbPath, { readonly: true });
+    try {
+        const rows = db.prepare(`
+      SELECT
+        id, source_type, source_id, target_type, target_id,
+        edge_type, confidence, method, sources_json, created_at, updated_at
+      FROM trace_edges
+      ORDER BY updated_at DESC
+    `).all();
+        const edges = rows.map((r) => ({
+            id: r.id,
+            source: { type: r.source_type, id: r.source_id },
+            target: { type: r.target_type, id: r.target_id },
+            type: r.edge_type,
+            confidence: r.confidence,
+            method: r.method,
+            sources: JSON.parse(r.sources_json || '[]'),
+            createdAt: r.created_at,
+            updatedAt: r.updated_at,
+        }));
+        const exportData = {
+            exportedAt: new Date().toISOString(),
+            totalEdges: edges.length,
+            edges,
+        };
+        await fs.writeFile(jsonPath, JSON.stringify(exportData, null, 2), 'utf8');
+        return { exported: edges.length, jsonPath };
+    }
     finally {
         db.close();
     }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "preflight-mcp",
-  "version": "0.1.9",
+  "version": "0.2.1",
   "description": "MCP server that creates evidence-based preflight bundles for GitHub repositories and library docs.",
   "type": "module",
   "license": "MIT",