Rifflux provides 5 tools with complete input schemas and reasonable descriptions. All tools have verb-noun naming (search_, get_, index_, reindex) and clear parameter types. However, descriptions are terse (avg 80 chars vs baseline 194), lack WHEN/WHY context, and output schemas are not documented. The reindex tool accepts mutually exclusive parameters (path vs paths) without explicit documentation of the constraint. Error handling is minimal, no recovery guidance or actionable error messages visible in the code.

• Document output schemas for all tools. E.g., search_rifflux should specify: returns array of {chunk_id, file_path, score, content, metadata}. get_chunk should specify: returns {chunk_id, file_path, content, metadata, related_chunks}.
• Expand descriptions to 100-200 chars with WHEN/WHY context. E.g., 'Get a specific indexed chunk by its stable ID. Use this after search_rifflux returns a chunk_id you want to examine in detail. Returns full content and metadata.'
• Add explicit constraint documentation for reindex: 'Exactly one of path or paths must be provided. path indexes a single file/directory; paths indexes multiple. Cannot specify both.'
• Implement structured error responses with recovery guidance. E.g., if index is missing, return: {error: 'Index not found', retryable: false, recovery: 'Call reindex() to build the index first.'}
• Add pagination support to search_rifflux: accept limit (1-100, default 20) and offset/cursor parameters; return total_count and next_cursor in response.

Inferred effective spec: 2025-06-18+.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry