deep-research

Automate multi-agent research by orchestrating parallel exploration, cross-checks, and citation audits through a phase-driven workflow. This protocol enforces URL fetching policies and validation gates to produce a verifiable final report with traceable sources saved on disk.

Coordinate deep research workflows using a structured protocol that moves through setup, scope, discovery, validation, and synthesis phases. This multi-agent coordination ensures thorough exploration and cross-checks, yielding source-grounded findings written as on-disk artifacts.

A phase-driven research workflow ensures source verification by enforcing strict phase gates between discovery, validation, and synthesis. It applies URL fetching policies and citation audits during cross-checks, guaranteeing that the final findings report contains only verified, traceable sources.

Yes, you can use multi-agent research for complex topic investigations and external lookups. The orchestration protocol automates exhaustive scoping and discovery across topics, applying cross-checks and citation audits to deliver thorough, source-grounded findings for any research request.

No external dependencies are required to run automated deep research orchestration. The workflow operates independently to orchestrate multi-agent phases from scope to synthesis, writing progress and final findings directly to on-disk artifacts for review.

Final findings and progress artifacts are stored directly on disk during a deep research run. The workflow writes progress tracking under docs/rpm/research and outputs a final structured findings report as report.md for review.