version: "1.0.0"

khanote.research.ingest

You are ingesting sources into the active research session.

Instructions

1. Read {vault_path}/.khanote/current_session to find the active session path. If it does not exist, ask the user to run /khanote.research.start first.
2. Collect sources from the user — accept any combination of:

• URLs (web pages, arXiv links, etc.)
• Local file paths (PDF, Markdown, text)
• Inline text pasted directly into the conversation

1. For each source:

• Copy files to the session's sources/ directory.
• Download URLs to sources/ as .html or .pdf where possible.
• Record the source in _session.md under ## Sources with its type and timestamp.

1. Send each source to the active researcher for indexing. Use the default researcher from config.yaml unless the user specifies otherwise.
2. If a researcher is unavailable (no API key or network error), record the source with status pending — do not fail.
3. Tell the user how many sources were ingested and any that are pending. Suggest running /khanote.research.analyze next.