researcher.default

Evidence collection from diverse web sources is performed by an autonomous agent that gathers facts, compares sources, and reports uncertainty to inform decision-making. It requires web search and fetch capabilities to retrieve and validate findings.

Source comparison and uncertainty reporting is handled by applying evidence-driven validation logic to gathered facts. The agent contrasts retrieved data, evaluates reliability, and outputs confidence notes indicating the level of uncertainty.

Validated research findings are persisted in a durable knowledge store designed to retain credible evidence for retrieval and reuse. This enables autonomous researchers to access previously validated facts without repeating web sourcing.

A sandboxed execution environment is required to safely gather facts and compare sources from the web. This isolation ensures the autonomous agent can execute web search and fetch capabilities without risking the host system.

Evidence gathering from both web sources and documents is supported by the agent's core functionality. It collects and organizes credible evidence across formats to enable evidence-based decision-making.

An evidence-driven autonomous research agent is needed when you must collect and organize credible evidence to inform decision-making. It applies to scenarios requiring autonomous fact-gathering, source comparison, and uncertainty reporting.