mphaxise Data Research
Use this skill when the user wants a research pipeline that extracts structured information into a durable tracker.
This port is adapted from garrytan/gbrain at commit b7e3005b5b3f1b54082f9c5990482ebf81a4a807.
Workflow
- Define the research target and tracker schema.
- Search the relevant sources:
- local files
- Gmail or other connectors
- web sources
- structured APIs when available
- Save raw evidence before summarizing.
- Extract structured fields.
- Deduplicate before adding to the tracker.
- Update the canonical tracker file or report.
Guardrails
- Save evidence before relying on memory or summary text.
- Do not append duplicate rows without checking.
- Keep the tracker schema stable across runs.