Build a period-end accrual schedule by computing each accrual, citing support, and drafting journal entries for controller approval. Use during month-end close; this drafts entries only and must not post them.
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。
Build and maintain a persistent, interlinked markdown wiki (second brain / research notebook / personal knowledge base) from raw sources. Use this skill when the user wants to ingest an article / paper / note / URL / PDF into their wiki, ask a question against their wiki, lint/health-check their wiki, or set up a new wiki from a folder. Triggers include phrases like "add this to my wiki", "ingest this", "file this article", "update my notes with this", "what does my wiki say about X", "health-check my wiki", "find stale/orphan pages", or any workflow involving a `raw/` source folder, `wiki/` markdown folder, `index.md`, or `log.md`. Compatible with Obsidian vaults.
インストール方法を見る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。
Treat the workspace as a persistent, compounding knowledge base. The human curates sources and asks questions; you do all the reading, summarizing, cross-referencing, filing, and bookkeeping.
When the task is a single clear sub-operation, prefer the specialized skills:
$wiki-ingest for filing one new source$wiki-query for answering from the wiki$wiki-lint for maintenance and health checks$wiki-bootstrap for initializing or restructuring the wikiUse this umbrella skill when the task spans multiple operations or when you need the overall workflow contract.
Interpreter exposes the note graph through interpreter_vault.
Use it before large wiki operations:
action="snapshot" to confirm workspace scale and graph shapeaction="search_notes" to find relevant notes by title, alias, or tagaction="note_context" to inspect backlinks, outgoing links, tags, and broken links for one noteaction="notes_with_tag" and action="list_tags" for tag-driven navigationaction="lint" for structural wiki health checksDo not manually rebuild the note graph with ad hoc grep when interpreter_vault already has the graph.
Default greenfield layout:
raw/ — immutable source documents (clipped articles, PDFs, notes, transcripts). Read-only. Never write to raw/.wiki/ — your output: summary, entity, concept, and synthesis pages. You own this layer entirely.index.md and log.md at the workspace root — navigation and history.If the workspace already has an established markdown/wiki structure, adopt it instead of forcing a parallel raw/ + wiki/ tree. Respect the user's existing folders, daily-note location, frontmatter conventions, and note naming. Only bootstrap the default layout when the workspace is actually empty or when the user explicitly asks for the raw/ + wiki/ structure.
wiki/)Every wiki page is a markdown file with YAML frontmatter. Use hyphen-case filenames matching the title.
---
title: Page Title
type: source | entity | concept | synthesis | comparison | overview
created: YYYY-MM-DD
updated: YYYY-MM-DD
sources: [raw/path/to/source1.md, raw/path/to/source2.pdf]
tags: [tag1, tag2]
---
[[wikilinks]] to every entity/concept it touches.[[wikilinks]].Link everything. Use [[Page Name]] wikilinks (Obsidian-compatible). If you reference something that doesn't have a page yet, link it anyway — orphan links surface what to create next.
When the user drops a new source into raw/ (or gives you a URL) and asks to ingest it:
wiki/sources/<title-slug>.md with frontmatter, citation, takeaways, quotes, and [[wikilinks]].> [!warning] Contradiction callout or inline ⚠️ note, citing both sources.index.md with a new entry under the appropriate section.log.md with the standard prefix format.A durable wiki workflow must already exist before these steps begin. If not, bootstrap first with $wiki-bootstrap.
A single ingest typically touches 5–15 wiki pages. That is expected and correct.
When the user asks a question of the wiki:
index.md first to find relevant pages.[[wikilinks]] as needed.raw/ sources). Use the form ([[Page Name]]) inline.When the user asks for a health-check:
[[wikilinks]] pointing to pages that don't exist yet.Report findings as a prioritized list with suggested next actions. Do not auto-fix unless the user asks.
index.md FormatContent-oriented catalog. Organized by section. One line per page:
# Index
## Synthesis
- [[Evolving thesis on X]] — current position across 12 sources
## Overviews
- [[Topic A]] — framing page for topic A
## Entities
- [[Person Name]] — short description
- [[Org Name]] — short description
## Concepts
- [[Concept Name]] — one-line hook
## Sources
- [[Source Title]] — author, date, 1-line summary
Update on every ingest. Keep one-line summaries tight (under ~80 chars).
log.md FormatChronological, append-only, grep-friendly. Every entry starts with the exact prefix:
## [YYYY-MM-DD] ingest | Source Title
- Created: [[Source Title]], [[New Entity]]
- Updated: [[Existing Concept]], [[Other Page]]
- Contradiction: [[Claim X]] (new source disagrees with [[Older Source]])
## [YYYY-MM-DD] query | "user's question verbatim or paraphrased"
- Filed as: [[Resulting Analysis Page]] (optional)
## [YYYY-MM-DD] lint | summary
- 3 orphans, 2 dangling links, 1 contradiction flagged
The prefix ## [YYYY-MM-DD] <op> | must be stable so grep "^## \[" log.md works.
raw/. Sources are immutable. If a source is wrong, note it on the corresponding wiki/sources/ page.index.md and log.md on every ingest. No silent changes.[[wikilinks]], not markdown [text](path.md) links, for inter-wiki references. Obsidian compatibility matters.raw/ source via its wiki/sources/ page.If the workspace is empty or clearly greenfield and has no raw/, wiki/, index.md, or log.md:
raw/ and wiki/ directories.index.md with empty section headers.log.md with a single entry: ## [YYYY-MM-DD] setup | wiki initialized.raw/ (or clip one with Obsidian Web Clipper).If the workspace already has a real markdown/wiki layout, skip bootstrap and work inside that layout.
This wiki is designed to work inside an Obsidian vault. [[wikilinks]], YAML frontmatter, and the default raw/ + wiki/ layout all render natively, but when the user already has a working vault structure, respect that structure instead of replacing it. The user can browse the graph view, use backlinks, and edit pages directly — you maintain consistency from the other side.
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
Build a period-end accrual schedule by computing each accrual, citing support, and drafting journal entries for controller approval. Use during month-end close; this drafts entries only and must not post them.
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。
Audit a spreadsheet for formula accuracy, errors, and common financial-model mistakes. Use for selected ranges, single sheets, or whole-workbook model checks including balance-sheet balance, cash tie-out, roll-forwards, and logic sanity.
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。
Use this skill when the user needs advanced Playwright control of an already running browser session through Interpreter's app-managed browser bridge and the `builtin-js-repl` `js_repl` tool, after simple browser page work cannot be handled by the unified `builtin-interpreter` browser page tools.
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。
Drive a native desktop app through Interpreter's builtin-cua-driver. Use get_app_state, click/type/scroll/drag, and verify by calling get_app_state again when the user asks to operate a real desktop app, browser chrome, native dialog, menu, secure prompt, file chooser, or hidden/background window.
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, inspect, and edit Word documents (`.docx`) with local code-execution tools. Use for document drafting, structured edits, tables, comments, and layout-sensitive Word deliverables.
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。
Design or revise Interpreter UI for non-developers with an ultra-minimal, utilitarian style inspired by the OpenAI website and ChatGPT, not the developer platform. Use when simplifying settings, onboarding, chat, forms, empty states, or helper copy; when removing extra containers, labels, or visual noise; when replacing technical wording with plain language; when clarifying hierarchy and making the main user action obvious; when making every visible label and status understandable to an average HR admin or other non-technical user; when hiding implementation details such as logs, IDs, endpoints, paths, or protocol names; or when a screen must respect the app's existing CSS variables and light/dark theme behavior.
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。