Spine provides 10 well-named, read-only architecture analysis tools with consistent verb-noun naming (architecture_*). All tools have descriptions and input schemas with typed parameters. However, descriptions are terse (10-50 chars), parameter descriptions lack actionable detail (format, constraints, examples), output schemas are not documented, and error handling guidance is absent. No tool annotations (readOnlyHint, etc.) despite all being read-only. Parameter relationships (e.g., 'from'/'source' alternatives in architecture_path) are undocumented. This is typical community-grade work: functional but not LLM-optimized.

• Expand tool descriptions to 50-150 chars. Include WHAT the tool does, WHEN to use it, and how it differs from similar tools. E.g., 'architecture_overview: Get a high-level summary of repository structure (modules, layers, dependencies). Use this first to understand the codebase. For detailed dependency paths, call architecture_path instead.'
• Add parameter descriptions with constraints and examples. E.g., 'mode: Indexing strategy. auto=detect language and index intelligently (default, recommended). full=index all files regardless of size. status_only=check index status without re-indexing.'
• Document output schemas for each tool. E.g., 'Returns: {results: [{id: string, name: string, type: enum[module|function|class], evidence: {file: string, line: number}}], total: number, nextCursor?: string}'
• Add readOnlyHint: true to all 10 tools to signal they are safe for speculative execution and do not modify state.
• Clarify mutually exclusive parameters. E.g., 'architecture_path: Use either (from, to) or (source, target), not both. Prefer (from, to) for new code.'