Vault Health
unlistedby Vault Health contributors
A local-first diagnostic for your vault: research integrity, source provenance, knowledge structure, and art & media health. Read-only; no AI, no network.
Vault Health
A local-first diagnostic for Obsidian vaults. It answers: what is structurally happening inside my knowledge system, and what deserves my attention?, not "how tidy are my Markdown files?"
- Research integrity: which research notes have recorded sources, which sources are never used, broken source links, duplicate sources, citations without metadata, disconnected research clusters.
- Art & Media: artwork records read through a configurable field mapping: artists, movements, periods, mediums, missing metadata, broken image links, unreadable dates.
- Research Workspaces: create named, persistent workspaces from notes you choose (references, never copies), each with a local graph, internal/external connection counts, clusters, associated sources and diagnostics. See docs/WORKSPACES.md.
- Research analysis: the automatic layer underneath: provenance of every research note across the vault, grouped into areas.
- Knowledge structure: broken links, orphan notes, isolated clusters, dead ends.
- Media: broken embeds, unreferenced PDFs, images, audio and video.
- Transparent scores: every number expands to affected, eligible, tolerance, weight and formula. A score is a diagnostic summary, not a verdict.
- Action Center: ranked findings, each with a reason, evidence, and Open / Inspect / Ignore / Mark reviewed. Housekeeping (singleton tags, oversized or empty notes) is secondary and collapsed.
Read-only. No AI. No network. No telemetry. It never edits, moves, renames or deletes anything. See docs/DATA.md, docs/SCORING.md, docs/RULES.md, docs/CONVENTIONS.md, docs/DEVELOPMENT.md.
Status (v0.1.0, MVP)
Implemented: scanner, convention engine, link structure, research integrity, Research Workspaces (create/rename/open/delete, add/remove notes, add connected, local graph, clusters, sources, diagnostics), research analysis, Art & Media (records, facets, metadata, images), basic media health, scoring, Action Center, settings. Planned next: art timeline, knowledge-graph analysis (bridges, centrality), duplicate/fragmentation analysis, art timeline, opt-in health history.
For plugin developers
Search results and similarity scores are powered by semantic analysis of your plugin's README. If your plugin isn't appearing for searches you'd expect, try updating your README to clearly describe your plugin's purpose, features, and use cases.