RivalSearchMCP provides 9 tools with reasonable coverage of research and content discovery workflows. Tools use clear verb-noun naming conventions (web_search, social_search, news_aggregation, etc.) that align with action intent. However, critical gaps exist: parameter descriptions are often present but lack specificity about constraints, enums, and allowed values. Output schemas are not explicitly documented in the source code, making it unclear what fields agents should expect. Error handling guidance is absent, no indication of recovery strategies or retryable vs fatal errors. Tool descriptions vary widely in quality (some are comprehensive multi-clause descriptions; others are brief). The fastmcp framework suggests automated schema generation, but validation against pattern:tool-chain and pattern:response-shaper reveals incomplete adherence.

• Document the output schema for every tool, including field names, types, and whether results are paginated. Example for web_search: { results: [{title, url, snippet, quality_score, date}], total_count, next_cursor }.
• Formalize enum constraints in JSON Schema input parameters. Replace 'mode' with explicit enum: { 'type': 'string', 'enum': ['topic', 'entity'], 'description': '...' }. This prevents hallucinated values.
• Split content_operations into six separate tools: retrieve_content, stream_content, analyze_content, extract_content, score_content, find_content_conflicts. Each should have a single clear responsibility.
• Split scientific_research into search_academic_papers and search_datasets. This enables agents to invoke the correct tool without deciding between operation modes.
• Add error handling guidance to every tool description. Example: 'On rate limit: tool will return error. Recommend waiting 60 seconds before retry. On no results: returns empty array (not an error).'
• Document pagination support. Specify whether tools support offset/limit or cursor-based pagination, and what fields (total_count, next_cursor, has_more) are returned.
• Add comparative guidance in tool descriptions. Example for web_search: 'Use this for broad web results. If you need to deeply crawl a specific website, use map_website instead.'
• Clarify mode/operation selection criteria. Example for research_topic: 'Use topic mode for open-ended research on subjects (e.g., 'machine learning trends'). Use entity mode for profiling specific people, companies, or projects (e.g., 'Anthropic the company').'

Inferred effective spec: 2025-06-18+.

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