Loading ToolBench

Reading the current corpus…

ToolsConnector MCP server · F · ToolBench

ToolsConnector

Enhanced MCP server for ToolsConnector with smart tool selection, multi-tenant serving, and schema optimization. Exposes connectors as MCP tools for AI assistants.

<=2025-11-25STDIO12 tools

47/100

Definition Quality52

Protocol Readiness32

Supportability47

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: Logging

Deprecated protocol patterns detected

Summary

12 tools with basic descriptions and partial schemas. No output schemas documented. Error handling absent. Naming is verb-noun compliant but descriptions are generic (avg ~60 chars, below 194-char baseline). No tool annotations (readOnlyHint/destructiveHint). Composition is sound (single responsibility per tool), but parameter documentation is severely incomplete, LLMs cannot infer what 'query', 'range', 'values' mean without descriptions. No pagination guidance for list tools. Security: credentials appear to be injected server-side (good), but no audit trail or permission gate patterns visible.

Tools (12)

gcalendar_create_eventwriteauth50/100

Create an event in Google Calendar

gcalendar_list_eventsread onlyauth50/100

List events from Google Calendar

gdrive_delete_filedestructiveauth50/100

Delete a file from Google Drive

gdrive_list_filesread onlyauth50/100

List files in Google Drive with optional search and filtering

gdrive_upload_filewriteauth50/100

Upload a file to Google Drive

gmail_list_emailsread onlyauthsource verified65/100

List emails from Gmail mailbox with optional query filtering

gsheets_get_valuesread onlyauth50/100

Read values from a Google Sheet

gsheets_update_valueswriteauth50/100

Issues (5)

CRITICALtool-description ↗

Parameter descriptions missing or minimal. 'query', 'range', 'values', 'buttons' lack actionable guidance. LLMs cannot infer format, constraints, or valid values without explicit descriptions.

slack_list_channelsgdrive_list_filesgsheets_get_valuesgsheets_update_valueswhatsapp_business_send_interactive_buttons

CRITICALresponse-shaper ↗

No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., file_id, event_id, channel_id). Breaks tool chaining.

gmail_list_emailsslack_list_channelsgdrive_list_filesgcalendar_list_eventsgsheets_get_values

HIGHpaginated-result ↗

List tools (gmail_list_emails, slack_list_channels, gdrive_list_files, gcalendar_list_events) lack pagination parameters (limit, offset, page_size). No guidance on result caps. Large result sets will blow context windows.

gmail_list_emailsslack_list_channelsgdrive_list_files

Recommendations

Add parameter descriptions to every input. For 'query' params, specify format (Gmail DSL, Drive DSL, etc.). For 'range', explain A1 notation. For 'values', clarify 2D array structure and type expectations.
Document output schemas for all tools. Specify which fields are returned, their types, and which IDs/references downstream tools need (e.g., file_id, event_id, channel_id).
Add pagination to list tools: accept limit (1 - 100, default 20) and offset/cursor parameters. Return total_count or next_cursor. Document result caps in descriptions.
Annotate tools with readOnlyHint (list/get tools), destructiveHint (delete tools), and idempotentHint (create/update tools). This enables agents to reason about safety without parsing text.
Wrap error responses with actionable guidance. E.g., 'Invalid spreadsheet_id: not found. Verify the ID is correct and you have access.' instead of raw 404.

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Logging (deprecated) - log to stderr or use OpenTelemetry

Score history

Overall score trend

First recorded score · v2 rubric

47/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	47	<=2025-11-25	v2