Mine recent Obsidian notes for creative, testable research questions by reading a compact cluster of recently updated notes, extracting tensions and transfer opportunities, and synthesizing ranked ideas grounded in the user's own note history. Use when the user wants idea generation, research questions, experiment concepts, thesis directions, or cross-note synthesis from their latest notes rather than generic brainstorming.
Turn recent note activity into research questions that are grounded in the user's actual work. Read a small bundle of recent notes, extract recurring motifs and unresolved tensions, then propose questions that are novel enough to matter and concrete enough to test.
7-21 days or roughly 8-20 recently modified Markdown
notes unless the user specifies a tighter scope..obsidian, Attachments, Templates, and other non-note
regions unless the user explicitly wants them.scripts/list_recent_notes.py when the vault is large or when you need a
deterministic candidate list before reading.For each selected note, extract only the information needed to reason:
projects, categories, tags, and related## Summary, ## Key sentence, ## Goals,
## Failure modes, ## Open questions, or their Korean equivalentsAvoid reading the entire vault by default. Keep the packet narrow enough that a cross-note synthesis is still coherent.
Look for question-worthy structure, not just interesting nouns.
Primary lenses:
When the notes are technical or fragmented, read references/idea-lenses.md and use those lenses explicitly.
Produce 3-7 candidate questions. For each one, include:
Prefer questions that are actionable in the user's real workflow. Do not fall back to generic templates like "Can we improve X with Y?" unless the notes truly support that jump.
Rank the questions by:
Mark each question as:
near-termmid-termspeculativeIf the user asks to save the output into the vault, create a standalone note