Server has 4 tools with complete input schemas and descriptions. All tools follow verb_noun naming (calculate_*, generate_*). Descriptions are domain-specific and explain the financial purpose (e.g., 'prevent the growth trap', 'prevent timing mismatch'). However, descriptions are terse (60-120 chars vs. baseline 194 chars), lack WHEN/WHY guidance for LLM selection, and output schemas are not documented. Parameters have types and descriptions but lack constraints (enums, ranges, formats). No error handling guidance beyond a single validation check in calculate_terminal_value. Tools are read-only and well-scoped, but lack recovery hints and actionable error messages.

• Document output schemas for all 4 tools. Example: calculate_wacc returns {cost_of_equity: float, wacc_base: float, final_discount_rate: float, weights: {debt: float, equity: float}}. Include in docstring or via JSON Schema.
• Expand descriptions to 150-250 chars with WHEN/WHY context. E.g., 'Calculate WACC to establish a defensible discount rate for DCF valuation. Call this first if you have cost of debt, tax rate, and capital structure. Returns final_discount_rate for use in calculate_terminal_value.'
• Add numeric constraints to parameter descriptions. E.g., 'risk_free_rate (0.0 - 0.1, typically 2 - 5% for developed markets)', 'beta (0.5 - 2.5, where 1.0 = market average)', 'tax_rate (0.0 - 0.5, corporate tax rate as decimal)'.
• Enhance error handling in calculate_wacc and calculate_eva. Validate total_capital > 0, wacc > 0, beginning_ic > 0. Return {error: str, guidance: str} on failure. E.g., 'total_capital must be > 0. Check that interest_bearing_debt + total_equity > 0.'
• Refactor generate_deal_strategy to return structured output: {strategy: 'OFFENSE'|'DEFENSE', rationale: str, negotiation_focus: str}. This enables LLMs to reason about strategy type and compose follow-up actions.