What problem does it solve? Research projects accumulate scattered notes, experiment logs, paper summaries, and results across a repository, making it hard to keep a durable, organized knowledge base in Obsidian without manual upkeep or MCP-based plugins. ## Core Features & Use Cases - Filesystem-first vault management: Detects, bootstraps, syncs, and archives a compact vault structure (Hub, Plan, Knowledge, Papers, Experiments, Results, Writing, Daily) via the deterministic project_kb.py script, with no MCP, API keys, or REST plugins required. - Agent-driven synthesis: Imports existing repositories by having an agent read key sources and write durable notes, following a Papers -> Experiments -> Results -> Writing promotion pipeline. - Knowledge CRUD rules: Enforces one canonical note per durable object, archive-by-default deletion, and bilingual (English/Chinese) note generation. - Use Case: A researcher with a machine learning repository asks to import it into Obsidian; the skill bootstraps the vault, synthesizes a project overview and experiment map, and keeps daily notes and results synchronized on subsequent turns. ## Quick Start Ask the assistant to import this research repository into my Obsidian vault and keep the project memory and daily notes synchronized.