researcher-deep-dive

To synthesize a multi-source research brief, provide a topic to research and the system pulls evidence from 3–5 distinct sources, labels each claim with a confidence level, and outputs a sourced Markdown brief with explicit citations.

Confidence tagging labels each research claim with a specific level: well-established, contested, emerging, or speculative. This mechanism provides an auditable provenance trail so readers can evaluate the reliability of synthesized findings.

Generating an auditable literature review requires providing your topic to the system, which then synthesizes findings into a TL;DR, explicitly lists open questions, and exports a Markdown brief with a provenance trail to .guild/runs/<run-id>/research/<slug>.md.

Yes, this approach supports product decisions needing auditable provenance by synthesizing evidence from 3–5 sources into a concise brief with explicit citations, confidence levels for each claim, and a documented provenance trail.

The best way to structure a multi-source investigation is to let the system generate a sourced brief that applies confidence labeling to each claim, provides a TL;DR, lists open questions, and exports the results to a Markdown file.

Yes, the research brief exports to a Markdown file located at .guild/runs/<run-id>/research/<slug>.md, enabling seamless integration with Guild workflows for downstream analysis and documentation.