@imazhar101/mcp-bigquery-server 1.0.0 → 1.0.2

package/dist/servers/bigquery/src/index.js

@@ -3,8 +3,8 @@ import { Server } from '@modelcontextprotocol/sdk/server/index.js';
 import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
 import { CallToolRequestSchema, ErrorCode, ListToolsRequestSchema, McpError, } from '@modelcontextprotocol/sdk/types.js';
 import { Logger } from '../../../shared/utils/logger.js';
-import { getEnvVar, getLogLevel, getOptionalEnvVar } from '../../../shared/utils/config.js';
-import { formatAsTable, formatAsJson } from '../../../shared/utils/formatter.js';
+import { getEnvVar, getLogLevel, getOptionalEnvVar, } from '../../../shared/utils/config.js';
+import { formatAsTable, formatAsJson, } from '../../../shared/utils/formatter.js';
 import { BigQueryService } from './services/bigquery-service.js';
 import { bigqueryTools } from './tools/index.js';
 class BigQueryServer {
@@ -16,7 +16,14 @@ class BigQueryServer {
         const projectId = getEnvVar('BIGQUERY_PROJECT_ID');
         const location = getOptionalEnvVar('BIGQUERY_LOCATION', 'US');
         const keyFilename = getOptionalEnvVar('BIGQUERY_KEY_FILE') || undefined;
-        this.service = new BigQueryService({ projectId, location, keyFilename }, this.logger);
+        const preventedFields = getOptionalEnvVar('BIGQUERY_PREVENTED_FIELDS')
+            .split(',')
+            .map((f) => f.trim())
+            .filter(Boolean);
+        if (preventedFields.length > 0) {
+            this.logger.info(`PII field filter active — blocking: ${preventedFields.join(', ')}`);
+        }
+        this.service = new BigQueryService({ projectId, location, keyFilename, preventedFields }, this.logger);
         this.server = new Server({ name: 'bigquery-server', version: '1.0.0' }, { capabilities: { tools: {} } });
         this.setupToolHandlers();
         this.setupErrorHandling();
@@ -42,26 +49,39 @@ class BigQueryServer {
             case 'query': {
                 const result = await this.service.executeQuery(args.sql, args.max_rows ?? 100, args.project_id);
                 if (!result.success) {
-                    return { content: [{ type: 'text', text: result.error }], isError: true };
+                    return {
+                        content: [{ type: 'text', text: result.error }],
+                        isError: true,
+                    };
                 }
-                const { rows, rowCount, schema, bytesProcessed, cacheHit } = result.data;
+                const { rows, rowCount, schema, bytesProcessed, cacheHit, strippedFields, } = result.data;
                 const headers = schema?.map((f) => f.name) ?? Object.keys(rows[0] ?? {});
                 const body = args.format === 'json'
                     ? formatAsJson(rows)
                     : formatAsTable(headers, rows);
                 const meta = [
                     `${rowCount} row${rowCount === 1 ? '' : 's'}`,
-                    bytesProcessed ? `${(Number(bytesProcessed) / 1e6).toFixed(2)} MB processed` : null,
+                    bytesProcessed
+                        ? `${(Number(bytesProcessed) / 1e6).toFixed(2)} MB processed`
+                        : null,
                     cacheHit ? 'cache hit' : null,
+                    strippedFields?.length
+                        ? `⚠️ ${strippedFields.length} PII field(s) redacted: ${strippedFields.join(', ')}`
+                        : null,
                 ]
                     .filter(Boolean)
                     .join(' · ');
-                return { content: [{ type: 'text', text: `${body}\n\n${meta}` }] };
+                return {
+                    content: [{ type: 'text', text: `${body}\n\n${meta}` }],
+                };
             }
             case 'list_datasets': {
                 const result = await this.service.listDatasets(args.project_id);
                 if (!result.success) {
-                    return { content: [{ type: 'text', text: result.error }], isError: true };
+                    return {
+                        content: [{ type: 'text', text: result.error }],
+                        isError: true,
+                    };
                 }
                 const rows = result.data.map((d) => ({
                     dataset: d.id,
@@ -71,14 +91,16 @@ class BigQueryServer {
                 }));
                 const text = rows.length === 0
                     ? 'No datasets found.'
-                    : formatAsTable(['dataset', 'location', 'created', 'description'], rows) +
-                        `\n\n${rows.length} dataset${rows.length === 1 ? '' : 's'}`;
+                    : formatAsTable(['dataset', 'location', 'created', 'description'], rows) + `\n\n${rows.length} dataset${rows.length === 1 ? '' : 's'}`;
                 return { content: [{ type: 'text', text }] };
             }
             case 'list_tables': {
                 const result = await this.service.listTables(args.dataset, args.project_id);
                 if (!result.success) {
-                    return { content: [{ type: 'text', text: result.error }], isError: true };
+                    return {
+                        content: [{ type: 'text', text: result.error }],
+                        isError: true,
+                    };
                 }
                 const rows = result.data.map((t) => ({
                     table: t.id,
@@ -98,7 +120,10 @@ class BigQueryServer {
             case 'get_table_schema': {
                 const result = await this.service.getTableSchema(args.dataset, args.table, args.project_id);
                 if (!result.success) {
-                    return { content: [{ type: 'text', text: result.error }], isError: true };
+                    return {
+                        content: [{ type: 'text', text: result.error }],
+                        isError: true,
+                    };
                 }
                 const rows = result.data.map((f) => ({
                     column: f.name,
@@ -115,13 +140,20 @@ class BigQueryServer {
             case 'preview_table': {
                 const result = await this.service.previewTable(args.dataset, args.table, args.limit ?? 10, args.project_id);
                 if (!result.success) {
-                    return { content: [{ type: 'text', text: result.error }], isError: true };
+                    return {
+                        content: [{ type: 'text', text: result.error }],
+                        isError: true,
+                    };
                 }
-                const { rows, rowCount, schema } = result.data;
+                const { rows, rowCount, schema, strippedFields } = result.data;
                 const headers = schema?.map((f) => f.name) ?? Object.keys(rows[0] ?? {});
+                const notice = strippedFields?.length
+                    ? `\n⚠️ ${strippedFields.length} PII field(s) redacted: ${strippedFields.join(', ')}`
+                    : '';
                 const text = rowCount === 0
                     ? `Table "${args.dataset}.${args.table}" is empty.`
-                    : formatAsTable(headers, rows) + `\n\n${rowCount} row${rowCount === 1 ? '' : 's'} previewed`;
+                    : formatAsTable(headers, rows) +
+                        `\n\n${rowCount} row${rowCount === 1 ? '' : 's'} previewed${notice}`;
                 return { content: [{ type: 'text', text }] };
             }
             default:

package/dist/servers/bigquery/src/services/bigquery-service.js

@@ -26,7 +26,6 @@ export class BigQueryService {
                 query: sql,
                 dryRun: true,
                 location: this.config.location,
-                defaultDataset: { projectId: billingProject },
             });
             const statementType = dryRunJob.metadata?.statistics?.query?.statementType ?? '';
             if (statementType !== 'SELECT') {
@@ -41,14 +40,15 @@ export class BigQueryService {
             const [job] = await this.client.createQueryJob({
                 query: sql,
                 location: this.config.location,
-                defaultDataset: { projectId: billingProject },
                 maximumBytesBilled: String(10 * 1024 * 1024 * 1024), // 10 GB safety cap
             });
             const [rows] = await job.getQueryResults({ maxResults: clampedRows });
             const jobMeta = job.metadata?.statistics?.query;
-            const schemaFields = job.metadata?.configuration?.query?.destinationTable
+            const schemaFields = job.metadata?.configuration?.query
+                ?.destinationTable
                 ? []
-                : ((await job.getMetadata())[0]?.statistics?.query?.schema?.fields ?? []);
+                : ((await job.getMetadata())[0]?.statistics?.query?.schema?.fields ??
+                    []);
             // Schema is more reliably fetched from the query response rows keys
             const headers = rows.length > 0 ? Object.keys(rows[0]) : [];
             const schema = schemaFields.length > 0
@@ -58,15 +58,24 @@ export class BigQueryService {
                     mode: f.mode ?? 'NULLABLE',
                     description: f.description,
                 }))
-                : headers.map((name) => ({ name, type: 'UNKNOWN', mode: 'NULLABLE' }));
+                : headers.map((name) => ({
+                    name,
+                    type: 'UNKNOWN',
+                    mode: 'NULLABLE',
+                }));
+            const { filteredRows, filteredSchema, strippedFields } = this.applyFieldFilter(rows, schema);
+            if (strippedFields.length > 0) {
+                this.logger.warn(`Stripped ${strippedFields.length} prevented field(s) from results: ${strippedFields.join(', ')}`);
+            }
             return {
                 success: true,
                 data: {
-                    rows,
-                    rowCount: rows.length,
-                    schema,
+                    rows: filteredRows,
+                    rowCount: filteredRows.length,
+                    schema: filteredSchema,
                     bytesProcessed,
                     cacheHit: jobMeta?.cacheHit ?? false,
+                    strippedFields: strippedFields.length > 0 ? strippedFields : undefined,
                 },
             };
         }
@@ -78,6 +87,34 @@ export class BigQueryService {
             };
         }
     }
+    applyFieldFilter(rows, schema) {
+        const prevented = this.config.preventedFields ?? [];
+        if (prevented.length === 0) {
+            return { filteredRows: rows, filteredSchema: schema, strippedFields: [] };
+        }
+        // Case-insensitive match against full field name or glob-style suffix (*_email, email_*)
+        const isBlocked = (name) => prevented.some((p) => name.toLowerCase() === p.toLowerCase());
+        const allFieldNames = schema.length > 0
+            ? schema.map((f) => f.name)
+            : rows.length > 0
+                ? Object.keys(rows[0])
+                : [];
+        const strippedFields = allFieldNames.filter(isBlocked);
+        if (strippedFields.length === 0) {
+            return { filteredRows: rows, filteredSchema: schema, strippedFields: [] };
+        }
+        const strippedSet = new Set(strippedFields.map((f) => f.toLowerCase()));
+        const filteredSchema = schema.filter((f) => !strippedSet.has(f.name.toLowerCase()));
+        const filteredRows = rows.map((row) => {
+            const clean = {};
+            for (const [k, v] of Object.entries(row)) {
+                if (!strippedSet.has(k.toLowerCase()))
+                    clean[k] = v;
+            }
+            return clean;
+        });
+        return { filteredRows, filteredSchema, strippedFields };
+    }
     async listDatasets(projectId) {
         const project = projectId ?? this.config.projectId;
         this.logger.info(`Listing datasets in project: ${project}`);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@imazhar101/mcp-bigquery-server",
-  "version": "1.0.0",
+  "version": "1.0.2",
   "description": "BigQuery MCP server — read-only query execution, schema exploration, dataset discovery",
   "type": "module",
   "main": "dist/servers/bigquery/src/index.js",