Loading ToolBench
MCP server for Apple Mail integration, providing tools to list accounts, search messages, manage mailboxes, send emails, handle drafts, manage rules, and work with templates.
24 tools with consistent verb-noun naming (send_email, list_accounts, delete_messages). All tools have descriptions (avg ~80 chars, within baseline 34-392). Input schemas present for all tools with typed parameters. However, descriptions are terse and lack LLM-optimized guidance on WHEN to use each tool vs alternatives. Output schemas are not documented, LLMs cannot predict response structure. Error handling is absent from visible code. Parameter descriptions exist but lack format constraints, ranges, or dependency hints. No confirmation/dry-run pattern for destructive operations (delete_messages, delete_rule, delete_template). Risk annotations present (READ_ONLY, WRITE, DESTRUCTIVE) but not surfaced in descriptions.
Create a new draft message.
Create a new mailbox in a specified account.
Create a new mail rule.
Create a new email template.
Delete messages permanently.
Delete a mail rule.
Delete an email template.
Forward a message to recipients.
Output schemas not documented. LLMs cannot predict response structure, fields, or types. Agents cannot plan downstream tool calls or extract required data reliably.
Destructive operations (delete_messages, delete_rule, delete_template) lack confirmation/dry-run pattern. No recovery guidance in descriptions. Agents can permanently delete data without safeguards.
Descriptions are terse (avg ~80 chars) and lack LLM-optimized guidance. Missing WHEN to use each tool vs alternatives (e.g., when to call search_messages vs get_messages). No dependency hints or prerequisites stated.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | <=2025-11-25 | v2 |
Retrieve draft messages from an account.
Retrieve detailed information about specific messages by their IDs.
List all email templates.
Retrieve the full conversation thread for a specified message.
List all configured email accounts in Apple Mail.
List all mailboxes in a specified email account.
List all mail rules configured in Apple Mail.
Reply to a message.
Save attachments from messages to the filesystem.
Search for messages in a specified account and mailbox with various filter criteria.
Send an email message.
Send an email message with file attachments.
Update an existing draft message.
Update message properties such as read status, flag status, or move to a different mailbox.
Update an existing mail rule.
Update an existing email template.
Parameter descriptions lack format constraints, ranges, or validation rules. E.g., date_from/date_to accept 'YYYY-MM-DD' but no validation guidance. flag_color accepts string but no enum of valid colors. LLMs will hallucinate invalid values.
No error handling guidance visible in tool definitions. Descriptions do not explain what errors can occur, how to recover, or what to try next. Agents will be stuck on failures.