ワンクリックでManusで任意のスキルを実行

lit-status

スター2

フォーク1

更新日2026年5月7日 23:16

Query the Obsidian literature library: counts by status, tags, projects; orphans; gaps; reading queue. Use /lit-map or "what's in my library", "what should I read next".

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

Manusで実行

ソース

The-Fish-Salmon

The-Fish-Salmon/claude-research-pack

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

Manusで実行

name: lit-status description: Query the Obsidian literature library: counts by status, tags, projects; orphans; gaps; reading queue. Use /lit-map or "what's in my library", "what should I read next".

Lit Status

A read-only skill that surfaces what's in 30_Literature/ and how it relates to the user's active projects.

What it answers

How many papers are in the library, broken down by status (unread / skimmed / read / cited / dismissed) and by year.
What tags / topics are present, and how concentrated.
What's been read recently (last 30 / 90 days, by added and by status change time).
What papers are cited in active project notes but missing from the library (so the user can capture them).
What papers in the library don't link to any project (orphaned reading).
Reading queue: unread papers, ranked by how often their citekey is referenced in active project notes.

Workflow

Use the obsidian MCP to enumerate notes in 30_Literature/. Read frontmatter for each (citekey, status, tags, projects, added, year, doi, arxiv).
Use obsidian MCP search to find citekey references across 10_Projects/, 60_MOCs/, 00_Inbox/. Build a citation count map.
Compose the requested report.

Modes

The skill supports several report shapes; the slash command or invocation passes a mode hint:

summary (default) -- one-paragraph overview + counts.
unread -- list of unread papers, ranked by project-reference count.
tags -- tag cloud + papers per tag.
gaps -- papers cited in projects but not captured.
orphans -- papers in library not linked to any project or MOC.
citation-map {citekey} -- for one paper, show which project notes reference it.
next-action -- "what should I read next?" Returns a single, prioritized next action in <=5 lines: the most-cited unread paper in the active project (or, if no active project, the most-recently-added unread paper), plus the exact slash command the user can run to act on it (typically /capture-paper or /copilot synthesize {citekey1} {citekey2}). Used by research-copilot for its Suggest move.

Constraints

Read-only. Never writes to the vault. If the user wants to update a paper's status, point them to the file path; don't edit it for them.
If the obsidian MCP is offline, fall back to filesystem Glob + Grep against the vault path from OBSIDIAN_VAULT_PATH.
Limit each report section to <=30 entries; if there are more, show top 30 and the total count.

Output style

Markdown tables for lists; one paragraph for summaries. Always include the count at the top of each table so the user knows whether they're seeing all or a subset.

Example output

## Library status (2026-04-27)

**Total**: 142 papers -- 61 read, 23 skimmed, 53 unread, 5 cited, 0 dismissed.
**Most-referenced unread** (top 5):
| Citekey | Title | References in active project notes |
|---|---|---|
| kim2023ionic | ... | 7 |
| ... |

**Tags** (top 8):
- ion-gated-transistor (38), reservoir-computing (29), edl (22), ...

**Gaps** (cited in 10_Projects/{active-slug} but not in library): 3
- 10.1038/s41928-... -> suggested: /capture-paper 10.1038/s41928-...
- ...

このリポジトリの他の Skills

同じリポジトリ

academic-deep-research

The-Fish-Salmon/claude-research-pack

Citation-disciplined literature pipeline backed by paper MCP servers (semantic-scholar, arxiv, paper-search). Use for lit review, fact-check, systematic review, scholarly research.

2026-05-072

deep-research

The-Fish-Salmon/claude-research-pack

Citation-disciplined literature pipeline backed by paper MCP servers (semantic-scholar, arxiv, paper-search). Use for /research, lit review, fact-check, systematic review.

2026-05-072

handoff

The-Fish-Salmon/claude-research-pack

Capture a structured handoff snapshot so the next chat can resume. Writes handoff_latest.md to the Claude Code memory dir and mirrors to the vault inbox. Use /handoff or "save context".

2026-05-072

ingest-pdf

The-Fish-Salmon/claude-research-pack

Ingest a local PDF (or folder of PDFs) into the Obsidian vault. Extracts DOI/title, resolves metadata, writes 30_Literature/CITEKEY.md. Use /ingest-pdf or "ingest this PDF".

2026-05-072

paper-capture

The-Fish-Salmon/claude-research-pack

Resolve a paper from a DOI, arXiv id, URL, or Author-Year-keyword into the Obsidian vault. Fetches metadata, downloads PDF, writes 30_Literature/CITEKEY.md. Use /capture-paper.

2026-05-072

research-copilot

The-Fish-Salmon/claude-research-pack

Interactive research co-pilot. Orients to current state, asks one clarifier per turn, suggests next moves, synthesizes papers, escalates. Use /copilot or "where is my research".

2026-05-072

Lit Status

A read-only skill that surfaces what's in 30_Literature/ and how it relates to the user's active projects.

What it answers

How many papers are in the library, broken down by status (unread / skimmed / read / cited / dismissed) and by year.

What tags / topics are present, and how concentrated.

What's been read recently (last 30 / 90 days, by added and by status change time).

What papers are cited in active project notes but missing from the library (so the user can capture them).

What papers in the library don't link to any project (orphaned reading).

Reading queue: unread papers, ranked by how often their citekey is referenced in active project notes.

Workflow

Use the obsidian MCP to enumerate notes in 30_Literature/. Read frontmatter for each (citekey, status, tags, projects, added, year, doi, arxiv).

Use obsidian MCP search to find citekey references across 10_Projects/, 60_MOCs/, 00_Inbox/. Build a citation count map.

Compose the requested report.

Modes

The skill supports several report shapes; the slash command or invocation passes a mode hint:

summary (default) -- one-paragraph overview + counts.

unread -- list of unread papers, ranked by project-reference count.

tags -- tag cloud + papers per tag.

gaps -- papers cited in projects but not captured.

orphans -- papers in library not linked to any project or MOC.

citation-map {citekey} -- for one paper, show which project notes reference it.

next-action -- "what should I read next?" Returns a single, prioritized next action in <=5 lines: the most-cited unread paper in the active project (or, if no active project, the most-recently-added unread paper), plus the exact slash command the user can run to act on it (typically /capture-paper or /copilot synthesize {citekey1} {citekey2}). Used by research-copilot for its Suggest move.

Constraints

Read-only. Never writes to the vault. If the user wants to update a paper's status, point them to the file path; don't edit it for them.

If the obsidian MCP is offline, fall back to filesystem Glob + Grep against the vault path from OBSIDIAN_VAULT_PATH.

Limit each report section to <=30 entries; if there are more, show top 30 and the total count.

Output style

Markdown tables for lists; one paragraph for summaries. Always include the count at the top of each table so the user knows whether they're seeing all or a subset.

Example output

## Library status (2026-04-27) **Total**: 142 papers -- 61 read, 23 skimmed, 53 unread, 5 cited, 0 dismissed. **Most-referenced unread** (top 5): | Citekey | Title | References in active project notes | |---|---|---| | kim2023ionic | ... | 7 | | ... | **Tags** (top 8): - ion-gated-transistor (38), reservoir-computing (29), edl (22), ... **Gaps** (cited in 10_Projects/{active-slug} but not in library): 3 - 10.1038/s41928-... -> suggested: /capture-paper 10.1038/s41928-... - ...