Ingest Claude Code conversation/session history into Obsidian as distilled knowledge. Use for importing or mining past Claude sessions, .claude data, project/session history, or audit logs.
日本語の概要は準備中です。原文の説明を表示しています。
Create dynamic, queryable Obsidian dashboard views using Bases or Dataview. Use for auto-updating tables, indexes, or structured views of vault content; not for ordinary question answering.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Two tools available: Obsidian Bases (native, GUI-driven, no plugin) and Dataview (community plugin, SQL-like, more powerful). Check which the user has and prefer Bases unless they ask for Dataview or need GROUP BY / computed columns.
Writing profile: Before drafting or rewriting natural-language Markdown, read and apply the Writing Profile Resolution section in llm-wiki/SKILL.md. Framework schema, provenance, safety, and operation-specific requirements take precedence.
Apply WRITING.md preferences only to optional Markdown dashboard prose; .base syntax remains unchanged.
llm-wiki/SKILL.md (inline @name override → walk up CWD for .env → global config → prompt setup). This gives OBSIDIAN_VAULT_PATH.$OBSIDIAN_VAULT_PATH/index.md to understand what categories and pages exist..base files)Bases are YAML files that define live views over vault notes. Native to Obsidian 1.8+, no plugin needed.
Top-level keys:
filters: # Global filter applied to all views (expression strings under and/or/not)
formulas: # Named computed properties — referenced as formula.<name>
properties: # Display config per property — sets displayName for column headers
summaries: # Aggregation formulas (e.g. mean, sum)
views: # Array of view definitions (required)
Each item in views::
views:
- type: table # table | list | cards | map
name: "View Name" # display label
limit: 50 # optional max rows
order: # column display order (list of property/formula names)
- file.name
- note.updated
groupBy: # grouping — goes INSIDE the view, NOT at top level
property: note.tags
direction: ASC # ASC | DESC
filters: # view-specific filter (merges with global filters)
and:
- 'note.status != "done"'
summaries:
formula.myFormula: Average
Filters use expression strings, not typed objects. Always wrap in and:, or:, or not: — a bare list causes a "may only have one of and/or/not keys" parse error.
# CORRECT
filters:
and:
- file.inFolder("concepts")
# WRONG — typed objects (parse error)
filters:
- type: folder
folder: concepts
Filters support nesting:
filters:
or:
- file.hasTag("book")
- and:
- file.inFolder("concepts")
- file.hasTag("research")
- not:
- file.hasTag("archived")
Different contexts use different naming — confirmed from Obsidian's auto-reformat behaviour:
| Context | Frontmatter field tags | File name | Formula |
|---|---|---|---|
properties: keys | note.tags | file.name | formula.<name> |
order: values | tags (bare) | file.name | formula.<name> |
groupBy.property: | tags (bare) | file.name | — |
filters: expressions | file.hasTag(...) / note.tags | file.name | formula.<name> |
formulas: expressions | note.tags, note.updated | file.name | — |
filters:
and:
- file.inFolder("concepts")
properties:
file.name:
displayName: Page
note.tags:
displayName: Tags
note.summary:
displayName: Summary
note.updated:
displayName: Updated
views:
- type: table
name: Table
order:
- file.name
- tags
- summary
- updated
filters:
and:
- file.inFolder("entities")
properties:
file.name:
displayName: Entity
note.title:
displayName: Full Name
note.tags:
displayName: Tags
note.summary:
displayName: Summary
views:
- type: cards
name: Cards
order:
- file.name
- title
- tags
- summary
When groupBy is set, omit that property from order: — it becomes the group header row and adding it as a column too causes duplication.
filters:
and:
- file.inFolder("concepts")
properties:
file.name:
displayName: Concept
note.summary:
displayName: Summary
note.updated:
displayName: Updated
views:
- type: table
name: By Domain
groupBy:
property: tags # bare property name, no note. prefix
direction: ASC
order:
- file.name # do NOT include tags here — already the group header
- summary
- updated
filters:
and:
- file.hasTag("machine-learning")
properties:
file.name:
displayName: Page
note.category:
displayName: Category
note.summary:
displayName: Summary
views:
- type: table
name: Table
order:
- file.name
- category
- summary
filters:
and:
- file.inFolder("projects")
- file.hasTag("active")
properties:
file.name:
displayName: Project
note.summary:
displayName: Summary
note.updated:
displayName: Last Updated
views:
- type: cards
name: Cards
order:
- file.name
- summary
- updated
filters:
or:
- file.inFolder("concepts")
- file.inFolder("entities")
properties:
file.name:
displayName: Page
note.category:
displayName: Category
note.updated:
displayName: Updated
views:
- type: table
name: Table
order:
- file.name
- category
- updated
filters:
and:
- file.inFolder("concepts")
formulas:
days_stale: "floor((now() - note.updated) / 86400000)"
properties:
file.name:
displayName: Page
note.updated:
displayName: Updated
formula.days_stale:
displayName: Days Stale
views:
- type: table
name: Stale
order:
- file.name
- updated
- formula.days_stale
| Expression | What it does |
|---|---|
file.inFolder("path") | Pages in that folder |
file.hasTag("tag") | Pages with that tag (no # prefix) |
file.hasLink("Note Name") | Pages linking to a note |
file.name == "note-name" | Exact filename match |
file.ext == "md" | Filter by extension |
note.propertyName | Any frontmatter property |
formula.formulaName | A named formula result |
now() | Current timestamp in ms |
On Obsidian UI-generated format: When Obsidian's GUI writes or reformats a
.basefile it may output a simplified shorthand with top-levelcolumns:,sort:, andview:keys instead of the canonical schema. That format also works — Obsidian accepts both. Manually authored files should use the canonical schema above.
Dataview uses a SQL-like query language inside ```dataview ``` code blocks in any note. More powerful than Bases for computed columns, GROUP BY, and cross-folder queries.
```dataview
TABLE
tags AS "Tags",
summary AS "Summary",
file.mtime AS "Last Modified"
FROM "concepts"
SORT file.mtime DESC
```
```dataview
TABLE WITHOUT ID
file.link AS "Entity",
tags AS "Tags",
summary AS "Summary"
FROM "entities"
SORT file.name ASC
```
rows. prefix after groupingAfter GROUP BY, individual file properties must be prefixed with rows. — otherwise the column is empty or errors.
```dataview
TABLE WITHOUT ID
rows.file.link AS "Concept",
rows.summary AS "Summary"
FROM "concepts"
GROUP BY tags[0] AS "Domain"
```
file.mtime for date mathAvoid choice(updated, date(updated), file.mtime) — mixed date formats in updated frontmatter cause arithmetic errors. file.mtime is always a valid DateTime.
```dataview
TABLE WITHOUT ID
file.link AS "Page",
category AS "Type",
file.mtime AS "Last Modified",
(date(today) - file.mtime).days + " days" AS "Age"
FROM "concepts" OR "entities" OR "projects"
WHERE file.name != file.folder
WHERE (date(today) - file.mtime).days > 30
SORT (date(today) - file.mtime).days DESC
```
```dataview
TABLE
summary AS "Summary",
file.mtime AS "Last Modified"
FROM "projects"
WHERE file.name != file.folder
SORT file.mtime DESC
```
| Clause | Usage |
|---|---|
FROM "folder" | All notes in folder |
FROM #tag | All notes with tag |
FROM "a" OR "b" | Union of two folders |
WHERE file.name != file.folder | Exclude folder index pages |
GROUP BY field AS "Label" | Group rows — use rows. for properties after this |
SORT field DESC | Sort direction |
file.link | Clickable wikilink |
file.mtime | Last modified time (always valid DateTime) |
(date(today) - file.mtime).days | Days since last modification |
Bases: Target path $OBSIDIAN_VAULT_PATH/_meta/<dashboard-name>.base
Dataview: Write queries directly into any .md note. A dedicated dashboard note at $OBSIDIAN_VAULT_PATH/_meta/dashboard.md works well for multi-section views.
Slug examples:
_meta/concepts-index.base_meta/recent-ingests.base_meta/projects-overview.base_meta/stale-pages.base_meta/dashboard.mdCreate _meta/ if it doesn't exist yet.
To embed a .base inside a note:
## Entities
![[_meta/entities-tracker.base]]
Ask before modifying an existing note.
Append to $OBSIDIAN_VAULT_PATH/log.md:
- [TIMESTAMP] WIKI_DASHBOARD name="<slug>" tool=bases|dataview view=<type> filter="<description>"
No manifest or index update needed — dashboards are live queries, not static pages.
| Dashboard | Best tool | What it shows |
|---|---|---|
| Content index | Bases or Dataview | All pages grouped by category, sorted by updated |
| Entity tracker | Bases (cards) | Entity pages as a visual card gallery |
| Concepts by domain | Dataview | Concepts grouped by first tag using GROUP BY |
| Ingestion log | Either | Pages sorted by created date |
| Stale content | Dataview | Pages not touched in 30+ days with day count |
| Project overview | Either | Project pages with last-sync date |
| Research tracker | Dataview | Synthesis pages tagged research |
and:/or:/not:, never typed objectsgroupBy goes inside the view definition — not as a top-level keyproperties: <name>: displayName: "...", not columns: [{title}]formulas: used for computed columns, referenced as formula.<name> in order/propertiesrows.property not bare propertyfile.mtime, not choice(updated, ...)_meta/ with a descriptive sluglog.md updated![[_meta/<name>.base]]) or open the dashboard noteQMD is a search index, not the source of truth. If $QMD_WIKI_COLLECTION is empty or unset, skip this step. Run it only after this skill has written or rewritten vault markdown. If QMD refresh fails, do not roll back the vault changes; report the QMD status separately.
Use $QMD_CLI if set; otherwise use qmd.
${QMD_CLI:-qmd} update
If the output says vectors are needed or embeddings may be stale, run:
${QMD_CLI:-qmd} embed
Verify the collection with either:
${QMD_CLI:-qmd} ls "$QMD_WIKI_COLLECTION"
or, when a specific page path is known:
${QMD_CLI:-qmd} get "qmd://$QMD_WIKI_COLLECTION/<page>.md" -l 5
Record one of:
QMD refreshed: update + embed + verifiedQMD refreshed: update only + verifiedQMD skipped: QMD_WIKI_COLLECTION unsetQMD skipped: qmd CLI unavailableQMD failed: <short error summary>まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Ingest Claude Code conversation/session history into Obsidian as distilled knowledge. Use for importing or mining past Claude sessions, .claude data, project/session history, or audit logs.
日本語の概要は準備中です。原文の説明を表示しています。
Map the load-bearing files, symbols, callers, and dependencies of an unfamiliar codebase before making changes. Use for architecture orientation, impact analysis, what-would-break questions, or finding callers. Uses CodeGraph when available and a dependency-free builtin fallback.
日本語の概要は準備中です。原文の説明を表示しています。
Ingest Codex CLI conversation/session history into Obsidian as distilled knowledge. Use for importing or mining ~/.codex sessions, rollout files, indexes, or transcript logs; use wiki-agent for targeted topic-only recall.
日本語の概要は準備中です。原文の説明を表示しています。
Ingest GitHub Copilot CLI/session history into Obsidian as distilled knowledge. Use for importing or mining Copilot history; not for general Copilot help, simple session search, or backup.
日本語の概要は準備中です。原文の説明を表示しています。
Find and add missing cross-references between wiki pages. Use when pages are disconnected or after large ingestion. This modifies pages to add links; use wiki-lint for report-oriented structural auditing.
日本語の概要は準備中です。原文の説明を表示しています。
Run or configure the daily wiki maintenance cycle: check source freshness, refresh the index and hot.md, and manage its scheduled 9 AM launchd/systemd/cron reminder. Use for daily or morning wiki refresh or scheduler setup and verification.
日本語の概要は準備中です。原文の説明を表示しています。