Knowledge Retrieval Agent

Overview

This agent is designed to act as an expert knowledge retrieval system. Its primary function is to answer complex evaluation questions by systematically querying a dedicated memory database, ensuring answers are grounded strictly in previously stored facts.

It follows strict protocols for searching, cross-referencing, and handling conflicting information to provide highly accurate and verifiable responses.

Capabilities

• Thorough Search: Uses provided tools (e.g., search_memory) with multiple queries to ensure comprehensive data gathering.
• Detail Retention: Prioritizes including exact numbers, names, dates, and specific details from the retrieved memory.
• Synthesis & Comparison: Capable of cross-referencing multiple entities by searching for each component separately before synthesizing a comparison.
• Temporal Awareness: Pays close attention to date ranges, preferring the most recent information when facts are known to change over time.
• Conflict Resolution: Explicitly identifies and reports discrepancies found in the memory rather than choosing one fact arbitrarily.
• Structured Output: Formats answers using numbered steps for procedures and clear comparative statements.

Example Use Cases

• Historical Analysis: Answering, "What were the primary causes of Event X according to records from 2018 versus 2022?" (Requires cross-referencing multiple time points).
• Procedural Guide: Generating step-by-step instructions for a process documented in memory. (Requires numbered list format).
• Fact Verification: Answering, "What is the current stated budget for Department A versus Department B?" (Requires specific data extraction and comparison).
• Gap Identification: If asked about a topic with no stored data, it will clearly state that the information could not be found in memory.