Find professional email addresses from name + company domain. Tests 15+ patterns against MX records with confidence scoring.
Single tool 'email_find_by_name' has a well-structured schema with all required parameters typed and described. The tool description is comprehensive (380+ chars) and explains use cases, output fields, and anti-patterns (what NOT to use it for). Input schema is complete with 3 required string parameters (domain, firstName, lastName) all properly described. Output structure is documented in the description with 6 fields (email, confidence, pattern, allCandidates, domain, mxValid). However, the output schema itself is not formally declared in the tool definition, only described in prose. Parameter descriptions are adequate but could be more prescriptive about format constraints (e.g., domain format, name length limits). No error handling guidance is visible in the tool definition.
Search for people/contacts matching criteria and find their professional email address from name and company domain. Alternative to Apollo people-search at 4x lower cost. Returns the most likely email with confidence score after testing 15+ patterns against MX records. 1. email (string) -- best matching email address found 2. confidence (number 0-100) -- likelihood the email is correct 3. pattern (string) -- the pattern that matched (e.g. "first.last", "flast", "first") 4. allCandidates (array) -- all tested patterns with individual scores 5. domain (string) -- company domain used 6. mxValid (boolean) -- whether domain has valid MX records Example output: {"email":"[email protected]","confidence":92,"pattern":"first.last","allCandidates":[{"email":"[email protected]","score":92},{"email":"[email protected]","score":75},{"email":"[email protected]","score":60}],"domain":"stripe.com","mxValid":true} Use this BEFORE sales outreach, cold emailing, or building prospect contact lists. Essential for searching for people/contacts and finding decision-maker emails when you only know their name and company. Drop-in replacement for Apollo people search. Do NOT use for email validation -- use email_verify_address instead. Do NOT use for company data -- use company_enrich_from_domain instead. Do NOT use for person data from email -- use person_enrich_from_email instead.
Output schema not formally declared. Tool description documents output fields (email, confidence, pattern, allCandidates, domain, mxValid) in prose and example JSON, but no structured output schema is present in the tool definition. LLMs cannot reliably parse prose-only output specs.
Parameter descriptions lack format constraints. 'domain' is described as 'Company domain (e.g. company.com)' but does not specify format (must be lowercase? no spaces? valid TLD?). 'firstName' and 'lastName' lack length limits or character restrictions. LLMs may pass invalid values.
No error handling guidance. Tool description does not explain what happens on failure (invalid domain, no MX records, no matching patterns). LLMs cannot determine if errors are retryable or how to recover.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 66 | 2025-06-18+ | v2 |
Tool description includes example output JSON in prose. While helpful for humans, LLMs tend to reuse example values literally (e.g., '[email protected]') in real calls. Use formal output schema instead.