feat: remove slides.find and sheets.find in favor of drive.search (#271)

* feat: remove slides.find and sheets.find in favor of drive.search

Closes #268

Remove the redundant find methods from SlidesService and SheetsService,
consolidating all file-finding through drive.search. This also drops
the Drive API dependency (drive_v3, getDriveClient, buildDriveSearchQuery,
MIME_TYPES) from both services entirely.

Changes:
- Remove slides.find and sheets.find tool registrations from index.ts
- Remove find methods and Drive API deps from SlidesService and SheetsService
- Remove associated tests and Drive API mocks
- Update WORKSPACE-Context.md with MIME type filter examples for all types
- Fix stale docs.find/docs.move references in Docs skill
- Create new Sheets and Slides skills with drive.search guidance

* refactor: move Sheets/Slides details from WORKSPACE-Context to skills

Remove the Docs/Sheets/Slides section (format selection, content handling)
since these are now covered by individual skills. Replace inline Sheets
nuances with a skill cross-reference, add Slides cross-reference.

* refactor: remove unused buildDriveSearchQuery and MIME_TYPES

Address PR review feedback:
- Remove buildDriveSearchQuery and MIME_TYPES from DriveQueryBuilder.ts
  (no longer used after removing slides.find and sheets.find)
- Keep only escapeQueryString (still used by DriveService)
- Rewrite DriveQueryBuilder tests to cover escapeQueryString only
- Add fullText contains note to Docs skill for consistency

* style: run prettier formatting

Author: allenhutchison

skills/docs/SKILL.md

@@ -216,12 +216,21 @@ All write tools (`writeText`, `replaceText`, `formatText`) accept an optional
 
 ### Finding Documents
 
-Use `docs.find` to search by title. Supports pagination with `pageToken`.
+Use `drive.search` with a document MIME type filter to find Google Docs:
+
+```
+drive.search({
+  query: "mimeType='application/vnd.google-apps.document' and name contains 'Weekly Report'"
+})
+```
+
+For full-text search across document content, use `fullText contains` instead of
+`name contains`.
 
 ### Moving Documents
 
-Use `docs.move` to move a document to a named folder. If multiple folders share
-the same name, the first match is used.
+Use `drive.moveFile` to move a document to a different folder. You can specify
+the destination by folder ID or folder name.
 
 ## Comments & Suggestions
 

skills/sheets/SKILL.md

@@ -0,0 +1,60 @@
+---
+name: sheets
+description: >
+  Activate this skill when the user wants to find, read, or analyze Google
+  Sheets spreadsheets. Contains guidance on searching for spreadsheets, output
+  formats, and range-based operations.
+---
+
+# Google Sheets Expert
+
+You are an expert at working with Google Sheets spreadsheets through the
+Workspace Extension tools. Follow these guidelines when helping users with
+spreadsheet tasks.
+
+## Finding Spreadsheets
+
+Use `drive.search` with a Sheets MIME type filter to find spreadsheets:
+
+```
+drive.search({
+  query: "mimeType='application/vnd.google-apps.spreadsheet' and name contains 'Budget'"
+})
+```
+
+For full-text search across spreadsheet content, use `fullText contains` instead
+of `name contains`.
+
+## Reading Data
+
+### Full Spreadsheet
+
+Use `sheets.getText` to read all sheets in a spreadsheet. Choose the output
+format based on the use case:
+
+- **text** (default): Human-readable with pipe-separated columns — good for
+  quick review
+- **csv**: Standard CSV format — good for data export and analysis
+- **json**: Structured JSON keyed by sheet name — good for programmatic
+  processing
+
+### Specific Range
+
+Use `sheets.getRange` with A1 notation to read a specific cell range:
+
+```
+sheets.getRange({
+  spreadsheetId: "spreadsheet-id",
+  range: "Sheet1!A1:D10"
+})
+```
+
+### Metadata
+
+Use `sheets.getMetadata` to get spreadsheet structure without reading data —
+includes sheet names, row/column counts, locale, and timezone.
+
+## ID Handling
+
+- All tools accept Google Drive URLs directly — no manual ID extraction needed
+- IDs and URLs are interchangeable in all `spreadsheetId` parameters

skills/slides/SKILL.md

@@ -0,0 +1,57 @@
+---
+name: slides
+description: >
+  Activate this skill when the user wants to find, read, or extract content from
+  Google Slides presentations. Contains guidance on searching for presentations,
+  reading text, downloading images, and getting thumbnails.
+---
+
+# Google Slides Expert
+
+You are an expert at working with Google Slides presentations through the
+Workspace Extension tools. Follow these guidelines when helping users with
+presentation tasks.
+
+## Finding Presentations
+
+Use `drive.search` with a Slides MIME type filter to find presentations:
+
+```
+drive.search({
+  query: "mimeType='application/vnd.google-apps.presentation' and name contains 'Quarterly Review'"
+})
+```
+
+For full-text search across presentation content, use `fullText contains`
+instead of `name contains`.
+
+## Reading Content
+
+### Text Extraction
+
+Use `slides.getText` to extract all text content from a presentation. Text is
+organized by slide with clear separators.
+
+### Metadata
+
+Use `slides.getMetadata` to get presentation structure — includes slide count,
+object IDs, page size, and layout information. Slide object IDs from metadata
+can be used with `slides.getSlideThumbnail`.
+
+## Downloading Images
+
+### All Images
+
+Use `slides.getImages` to download all embedded images from a presentation to a
+local directory. Requires an **absolute path** for the output directory.
+
+### Slide Thumbnails
+
+Use `slides.getSlideThumbnail` to download a thumbnail of a specific slide.
+Requires the slide's `objectId` (from `slides.getMetadata` or `slides.getText`)
+and an **absolute path** for the output file.
+
+## ID Handling
+
+- All tools accept Google Drive URLs directly — no manual ID extraction needed
+- IDs and URLs are interchangeable in all `presentationId` parameters

workspace-server/WORKSPACE-Context.md

@@ -82,24 +82,15 @@ When creating documents in specific folders:
 2. Move it to the target folder with `drive.moveFile`
 3. Confirm successful completion
 
-To find Google Docs, use `drive.search` with a document MIME type filter rather
-than searching by name alone.
+To find Google Docs, Sheets, or Slides, use `drive.search` with a MIME type
+filter rather than searching by name alone. Example MIME type queries:
 
-## 📄 Docs, Sheets, and Slides
-
-### Format Selection (Sheets)
-
-Choose output format based on use case:
-
-- **text**: Human-readable, good for quick review
-- **csv**: Data export, analysis in other tools
-- **json**: Programmatic processing, structured data
-
-### Content Handling
-
-- Docs/Sheets/Slides tools accept URLs directly - no ID extraction needed
-- Use markdown for initial document creation when appropriate
-- Preserve formatting when reading/modifying content
+- Docs:
+  `mimeType='application/vnd.google-apps.document' and name contains 'query'`
+- Sheets:
+  `mimeType='application/vnd.google-apps.spreadsheet' and name contains 'query'`
+- Slides:
+  `mimeType='application/vnd.google-apps.presentation' and name contains 'query'`
 
 ## 🚫 Common Pitfalls to Avoid
 
@@ -186,9 +177,13 @@ Choose output format based on use case:
 
 ### Google Sheets
 
-- Multiple output formats available
-- Range-based operations with A1 notation
-- Metadata includes sheet structure information
+- See the **Sheets skill** for detailed guidance on finding spreadsheets, output
+  format selection, and range-based operations.
+
+### Google Slides
+
+- See the **Slides skill** for detailed guidance on finding presentations, text
+  extraction, image downloads, and slide thumbnails.
 
 ### Google Calendar
 

workspace-server/src/__tests__/services/SheetsService.test.ts

@@ -24,7 +24,6 @@ describe('SheetsService', () => {
   let sheetsService: SheetsService;
   let mockAuthManager: jest.Mocked<AuthManager>;
   let mockSheetsAPI: any;
-  let mockDriveAPI: any;
 
   beforeEach(() => {
     // Clear all mocks before each test
@@ -45,15 +44,8 @@ describe('SheetsService', () => {
       },
     };
 
-    mockDriveAPI = {
-      files: {
-        list: jest.fn(),
-      },
-    };
-
     // Mock the google constructors
     (google.sheets as jest.Mock) = jest.fn().mockReturnValue(mockSheetsAPI);
-    (google.drive as jest.Mock) = jest.fn().mockReturnValue(mockDriveAPI);
 
     // Create SheetsService instance
     sheetsService = new SheetsService(mockAuthManager);
@@ -306,60 +298,6 @@ describe('SheetsService', () => {
     });
   });
 
-  describe('find', () => {
-    it('should find spreadsheets by query', async () => {
-      const mockResponse = {
-        data: {
-          files: [
-            { id: 'sheet1', name: 'Spreadsheet 1' },
-            { id: 'sheet2', name: 'Spreadsheet 2' },
-          ],
-          nextPageToken: 'next-token',
-        },
-      };
-
-      mockDriveAPI.files.list.mockResolvedValue(mockResponse);
-
-      const result = await sheetsService.find({ query: 'budget' });
-      const response = JSON.parse(result.content[0].text);
-
-      expect(mockDriveAPI.files.list).toHaveBeenCalledWith({
-        pageSize: 10,
-        fields: 'nextPageToken, files(id, name)',
-        q: "mimeType='application/vnd.google-apps.spreadsheet' and fullText contains 'budget'",
-        pageToken: undefined,
-        supportsAllDrives: true,
-        includeItemsFromAllDrives: true,
-      });
-
-      expect(response.files).toHaveLength(2);
-      expect(response.files[0].name).toBe('Spreadsheet 1');
-      expect(response.nextPageToken).toBe('next-token');
-    });
-
-    it('should handle title-specific searches', async () => {
-      const mockResponse = {
-        data: {
-          files: [{ id: 'sheet1', name: 'Q4 Budget' }],
-        },
-      };
-
-      mockDriveAPI.files.list.mockResolvedValue(mockResponse);
-
-      const result = await sheetsService.find({ query: 'title:"Q4 Budget"' });
-      const response = JSON.parse(result.content[0].text);
-
-      expect(mockDriveAPI.files.list).toHaveBeenCalledWith(
-        expect.objectContaining({
-          q: "mimeType='application/vnd.google-apps.spreadsheet' and name contains 'Q4 Budget'",
-        }),
-      );
-
-      expect(response.files).toHaveLength(1);
-      expect(response.files[0].name).toBe('Q4 Budget');
-    });
-  });
-
   describe('getMetadata', () => {
     it('should retrieve spreadsheet metadata', async () => {
       const mockSpreadsheet = {

workspace-server/src/__tests__/services/SlidesService.test.ts

@@ -44,7 +44,6 @@ describe('SlidesService', () => {
   let slidesService: SlidesService;
   let mockAuthManager: jest.Mocked<AuthManager>;
   let mockSlidesAPI: any;
-  let mockDriveAPI: any;
 
   beforeEach(() => {
     // Clear all mocks before each test
@@ -62,15 +61,8 @@ describe('SlidesService', () => {
       },
     };
 
-    mockDriveAPI = {
-      files: {
-        list: jest.fn(),
-      },
-    };
-
     // Mock the google constructors
     (google.slides as jest.Mock) = jest.fn().mockReturnValue(mockSlidesAPI);
-    (google.drive as jest.Mock) = jest.fn().mockReturnValue(mockDriveAPI);
 
     // Create SlidesService instance
     slidesService = new SlidesService(mockAuthManager);
@@ -195,62 +187,6 @@ describe('SlidesService', () => {
     });
   });
 
-  describe('find', () => {
-    it('should find presentations by query', async () => {
-      const mockResponse = {
-        data: {
-          files: [
-            { id: 'pres1', name: 'Presentation 1' },
-            { id: 'pres2', name: 'Presentation 2' },
-          ],
-          nextPageToken: 'next-token',
-        },
-      };
-
-      mockDriveAPI.files.list.mockResolvedValue(mockResponse);
-
-      const result = await slidesService.find({ query: 'test query' });
-      const response = JSON.parse(result.content[0].text);
-
-      expect(mockDriveAPI.files.list).toHaveBeenCalledWith({
-        pageSize: 10,
-        fields: 'nextPageToken, files(id, name)',
-        q: "mimeType='application/vnd.google-apps.presentation' and fullText contains 'test query'",
-        pageToken: undefined,
-        supportsAllDrives: true,
-        includeItemsFromAllDrives: true,
-      });
-
-      expect(response.files).toHaveLength(2);
-      expect(response.files[0].name).toBe('Presentation 1');
-      expect(response.nextPageToken).toBe('next-token');
-    });
-
-    it('should handle title-specific searches', async () => {
-      const mockResponse = {
-        data: {
-          files: [{ id: 'pres1', name: 'Specific Title' }],
-        },
-      };
-
-      mockDriveAPI.files.list.mockResolvedValue(mockResponse);
-
-      const result = await slidesService.find({
-        query: 'title:"Specific Title"',
-      });
-      const response = JSON.parse(result.content[0].text);
-
-      expect(mockDriveAPI.files.list).toHaveBeenCalledWith(
-        expect.objectContaining({
-          q: "mimeType='application/vnd.google-apps.presentation' and name contains 'Specific Title'",
-        }),
-      );
-
-      expect(response.files).toHaveLength(1);
-      expect(response.files[0].name).toBe('Specific Title');
-    });
-  });
-
   describe('getMetadata', () => {
     it('should retrieve presentation metadata', async () => {
       const mockPresentation = {