What problem does it solve?
Finding and combining information scattered across a personal knowledge base is slow and error-prone, and answers without citations are hard to trust. This Skill provides a structured research workflow that decomposes questions, runs targeted knowledge base searches, and produces cited answers without fabricating sources.
Core Features & Use Cases
- Question Decomposition: Breaks a broad research question into 2-4 specific, citation-worthy sub-questions before searching.
- Targeted KB Search: Runs focused searches using terms the user actually wrote, then opens the most relevant documents fully with read_file and uses grep to locate supporting passages in long files.
- Citation Discipline: Every claim carries the [n] label from the source passage, and missing information is explicitly flagged as "Not found in your knowledge base" instead of being invented.
- Use Case: Ask "summarize what we've discussed about project X across all my notes" and receive a short synthesized answer with inline citations drawn from multiple documents.
Quick Start
Ask the agent to research a topic inside your knowledge base, for example: find everything in my knowledge base about the onboarding redesign and summarize it with citations.
