Knowledge Synthesis

What it does

This skill takes raw search results from multiple sources—such as chat logs, emails, cloud storage documents, and project trackers—and transforms them into a coherent and trustworthy summary. It identifies overlapping information, merges duplicates while preserving unique details, and provides clear attribution to the original sources. The goal is to provide a concise answer that synthesizes insights from disparate data points.

When to use it

• When you need to quickly understand a decision or outcome based on information spread across various communication channels.
• To consolidate findings from research involving multiple documents and discussions.
• For summarizing project progress when updates are scattered throughout different tools.
• To create a unified view of a topic where the same information is repeated in different formats (e.g., chat, email, documentation).

Key capabilities

• Cross-Source Deduplication: Identifies and merges duplicate information from various sources based on content similarity, sender/author, timestamps, entity references, and source citations.
• Prioritized Merging: Combines information by prioritizing completeness, authority (official documents over chats), and recency.
• Citation & Source Attribution: Provides inline citations within the synthesized answer and a complete list of sources at the end.
• Narrative Synthesis: Combines information into a single, cohesive narrative.

Example prompts

• "Synthesize these search results: [list of chat logs, emails, documents, project tracker entries]"
• “Create a summary from these sources about the API decision: [list of sources]"
• "Combine this information and provide attribution: [paste raw data from multiple sources]"

Tips & gotchas

• The skill prioritizes complete versions of information. Ensure all relevant sources are included for best results.
• Conflicting viewpoints or evolving information between sources will be kept as separate items, rather than merged.
• Pay attention to the source attribution; it's crucial for verifying the synthesized information.