Single tool with basic schema and description. Tool name follows verb_noun pattern (send-transactional-email), which is good. Description is 50 chars, within acceptable range but minimal. Input schema is present with proper types (array, string) and descriptions for all three parameters. However, output schema is completely undocumented, the tool returns raw API response text without structure. No error handling guidance, no parameter constraints (e.g., email format validation), and no indication of side effects or idempotency. Error handling catches exceptions but returns unstructured text rather than actionable recovery guidance. Missing: output schema documentation, parameter format constraints, error classification, and confirmation pattern for this destructive operation (sends email).
Sends a transactional email via Dotdigital API.
Output schema completely undocumented. Tool returns raw API response text without structure. LLM cannot parse response fields or plan downstream operations.
No error handling guidance. Catch block returns generic error text. LLM cannot determine if error is retryable, user-fixable, or fatal.
Destructive operation (sends email) lacks confirmation or dry-run pattern. No idempotency guarantee. Agents may accidentally send duplicate emails on retry.
Parameter 'toAddresses' accepts any string array without email format validation. LLM may pass invalid addresses; tool will fail at API call time with unhelpful error.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
No parameter descriptions explain expected format, length limits, or constraints. 'subject' and 'content' lack guidance on max length, encoding, or special character handling.