Two tools with minimal schema rigor and weak descriptions. about_me has empty input schema ({}), violating schema requirements. contact has proper parameter types but descriptions lack actionable detail. Both descriptions are generic and under 200 chars. No error handling guidance, no output schema documentation, no parameter constraints (enums, ranges). Tool names lack clear verb-noun structure ('about_me' is vague; 'contact' is ambiguous, send? receive?). No security considerations documented despite contact being a WRITE operation that sends external messages.

• Rename about_me → get_personal_info and contact → send_contact_message. Use verb_noun convention so LLM infers intent from name alone.
• Document output schemas explicitly. E.g., about_me returns {name: string, email: string, skills: [string], ...}. contact returns {status: 'success'|'error', message: string, timestamp: ISO8601}.
• Add parameter descriptions for contact: sender_email should state 'Valid email address of message sender (required, no examples)'. subject: 'Message subject line (1-200 chars)'. message: 'Message body (1-5000 chars)'.
• Add error handling guidance to contact description: 'On webhook failure, returns {status: error, message: ...}. Retry with exponential backoff. If webhook URL is invalid, contact server admin.'
• Move webhook_url to environment variable (CONTACT_WEBHOOK_URL) and validate it at server startup. Sanitize sender_email against email injection. Add rate limiting (max 5 messages per minute per sender).