Both tools lack input schemas entirely (empty {} objects), descriptions are minimal (under 20 chars), and no parameter documentation exists. Tool names follow verb_noun convention (get_me, list_prospects) but descriptions provide no context for LLM selection, prerequisites, or return structure. No output schemas documented. No error handling guidance. This server fails multiple hard scoring rules: missing input schemas (→ schema score 0), descriptions under 20 chars (→ description score 0-20).

• Add explicit input schemas to both tools. Even if they accept no parameters, document this: {"type": "object", "properties": {}, "required": []}. If future versions add filters (e.g., list_prospects with status or limit), add them now with descriptions.
• Expand descriptions to 50 - 200 characters. Example: 'get_me: Retrieve the authenticated user's profile including ID, email, name, and account status. Call this first to verify authentication and get your user_id for other calls.' Example: 'list_prospects: List all prospects in your account with name, email, status, and last_contact_date. Returns up to 50 results; use limit and offset for pagination.'
• Document output schemas. Specify what fields each tool returns (e.g., get_me returns {user_id, email, name, account_status}; list_prospects returns {prospects: [{id, name, email, status, last_contact_date}], total, limit, offset}).
• Add pagination to list_prospects: accept limit (1 - 100, default 20) and offset (default 0) parameters; return total count and next_offset so agents can fetch all results without context explosion.
• Add error handling guidance: 'If authentication fails, verify LINKREACHER_API_KEY is set. If rate limited, retry after 60 seconds. If no prospects found, check account status or filters.'