Two tools with clear, actionable descriptions (194 and 156 chars respectively, within baseline 34-392 range). Both have input schemas with typed parameters and descriptions. However, output schemas are completely undocumented, LLMs cannot plan downstream calls or extract structured data. No error handling guidance, no pagination support despite search tool returning potentially large result sets, and no tool annotations (readOnlyHint/destructiveHint). Naming is verb-forward (sec_regulation, sec_regulations_search) but generic 'sec_regulation' lacks specificity about what 'regulation' means (full text? metadata? citations?). Parameter descriptions are adequate but lack format constraints (e.g., citation format examples are in tool description, not parameter description). Missing output schema documentation is a critical gap for agent composition.

• Document output schema for both tools. Example: sec_regulation returns {citation: string, title: string, text: string, effective_date: string, amendments: string[]}. sec_regulations_search returns {results: [{citation: string, title: string, excerpt: string, relevance_score: number}], total_count: number, next_cursor?: string}.
• Add pagination to sec_regulations_search: limit (1-100, default 20), offset or cursor, and total_count in response. Include guidance: 'Returns max 20 results per call; use offset to fetch additional pages.'
• Add error handling guidance to both tool descriptions. Example: 'If citation not found, returns error with available similar citations. If network timeout, retry after 5 seconds.'
• Declare tool annotations in McpToolDefinition: add readOnlyHint: true to both tools to signal they are safe to call without user confirmation.
• Move citation format constraints into parameter schema as enum or pattern. Example: description should say 'Format: CFR section (e.g., 240.10b-5), rule name (e.g., Rule 10b-5), part number (e.g., part 240), or section with subsections (e.g., 230.506(b))'.