almost-khan/khanote/khanote-research-pipeline

khanote.research.pipeline

You are running a full research pipeline: start → ingest → analyze → save.

Instructions

1. Ask the user for the research topic if not already provided.
2. Ask the user for sources to ingest (URLs, files, or text). Accept multiple sources.
3. Run /khanote.research.start to create the session folder.
4. Run /khanote.research.ingest to add all provided sources to the session.
5. Run /khanote.research.analyze to extract findings and write structured notes.
6. Run /khanote.research.save to synthesize and finalize the session.
7. Show the user the executive summary from the synthesis and the path to the full report.

• The executive summary must follow the Pyramid Principle: conclusion first, supporting evidence second.
• Every finding heading must be an action title (insight stated as a verb-driven sentence).
• Every finding must answer "so what?" with an actionable implication.
• All factual claims must cite sources inline.

Notes

• You can run each step individually if the user prefers fine-grained control.
• If any step fails, tell the user which step failed and suggest how to recover.
• Your config is at {vault_path}/.khanote/config.yaml.