research-messages-search

Search conversations by query using the research-messages-search Skill to locate discussions about papers, topics, or findings across multiple sessions. It performs hybrid and semantic searches on your conversation history, returning relevant matches with optional time and context limits.

Yes, context expansion retrieves surrounding messages around selected results so you see full discussion threads. This helps you understand the complete conversation flow and relationships between research topics without jumping between fragments.

Use semantic search to discover conversations spanning different research topics and sessions. The Skill indexes your full conversation history and surfaces cross-session connections, letting you map how topics evolved and relate across time.

No external dependencies are required. The Skill runs via npx tsx on Node.js and searches your existing conversation history directly, making it accessible for any research workflow without additional configuration or data preparation.

Semantic search understands meaning and context beyond exact keywords, matching discussions about related concepts and paper themes even when wording differs. This surfaces relevant research conversations keyword matching would miss.

Search performance depends on conversation volume and query complexity. Use time limits and context constraints to narrow results when searching across many sessions, and refine queries if results feel too broad or unfocused.