REA
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
📼 Declarative AI agent environment manager, written in Rust
作者の方は、このコードを README に貼れます。バッジはこのページにリンクします。
[](https://mcp.popnx.com/en/skills/pivoshenko-kasetto)同じ作業で AI をサービスにつなぐ MCP
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
Codebase intelligence for TypeScript and JavaScript, exposed as structured MCP analysis tools.
MCP server for terminal commands, file operations, and process management
同じ種類の作業。Skill 一覧と同じ並び
| これ kasetto | rea | andrej-karpathy-skills | graphify | MoneyPrinterTurbo | |
|---|---|---|---|---|---|
| 作者 | pivoshenko | morluto | multica-ai | Graphify-Labs | harry0703 |
| Skill 数 | 1 | 1 | 1 | 1 | 1 |
| GitHub スター | 219 | 467,235 | 234,125 | 149,873 | 144,332 |
| 最終コード更新 | 2026年11月15日 | 2026年11月18日 | 2026年4月20日 | 2026年11月18日 | 2026年11月18日 |
| ライセンス | – | MIT | – | Apache-2.0 | MIT |
pivoshenko による 1 個の Skill(指示セット)。AI が読んでそのとおりに作業します · MCP server ではありません
MCP = 手
AI を外部のサービスやデータにつなぎます(例:Notion を読む、請求書を作る)。プログラムやサーバーが動いている必要があります。
Skill = 手順書
作業のやり方を AI に教えるテキストファイル。どのサービスにもつながりません。
重なる部分もあります。ツールを呼ぶスクリプト付きの Skill や、知識だけを渡す MCP もあります。やりたい作業から選んでください。
Claude CodeCodexCursorGemini CLIGitHub CopilotWindsurfOpenCodeKiro
提供元の README と repo のフォルダから判断
GitHub account pivoshenko (person), 8 years old
★ 219 stars, few users so far · code last changed 2026年11月15日
2 scripts that can run on a computer: scripts/collect.py, scripts/render.py
No internet calls found in the scripts
No hooks or writes to machine settings found
Checked automatically from the repo files at commit 687d711 (read only, nothing was run). Not a safety guarantee: read it and decide for yourself
まず:どの AI を使っていますか
使っているアプリを選ぶと、そのアプリでのインストール方法だけが表示されます
This skill has no license, so we do not hand out a ZIP from our site. Download it from GitHub and arrange the folder yourself, or use "Let the AI install it" above, which is easier
Download from GitHub (whole repo) ↗
Unzip it, then zip only the skills/kst-ai-assets-usage folder again. The ZIP must have a kst-ai-assets-usage folder with SKILL.md at the top level, no loose files
+Create skillUpload a skillpick kst-ai-assets-usage.zip
SettingsCapabilitiesCode execution and file creation
Turn on Code execution and file creation first (newer menus: Create and edit files) · personal plans: Settings → Capabilities · Team/Enterprise: an admin turns it on under Organization settings → Plugins & skills · every plan, including Free
Upload once; it works in claude.ai and Claude Desktop, and in recent Claude Code signed in with a Claude account (not with an API key)
Cowork and cloud sessions use the skills turned on in your claude.ai account, not ~/.claude/skills on your computer (Claude Desktop scheduled tasks run locally and still load it)
Checked against official docs 2026年11月12日 · Source: support.claude.com, support.claude.com, code.claude.com
この文をコピーして Claude Code のチャットに貼り付けます
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Claude Code を開き直し、/kst-ai-assets-usage と入力するか関連する作業を頼むと Claude が自動で使います
Checked against official docs 2026年11月12日 · Source: code.claude.com
Skill は AI が従う指示です。入れる前にファイルを読み、信頼できる提供元のものだけを入れてください。
Mac:Command+スペースで Terminal と入力 · Windows:下のコマンドは PowerShell では動きません。上の「AI にインストールしてもらう」を使ってください
repo を一時フォルダにダウンロードし、SKILL.md のあるフォルダだけをコピーします
コマンドを確認中 · 確認が通るとコピーボタンが出ます
ChatGPT デスクトップアプリには Codex が組み込まれていて、Codex は ~/.agents/skills フォルダから Skill を読み込みます
ChatGPT デスクトップアプリで Codex を開き、この文をチャットに貼り付けます
The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
kst-ai-assets-usage フォルダを ~/.agents/skills/kst-ai-assets-usage に置きます
~ はホームフォルダ($HOME)です
Skill は AI が従う指示です。入れる前にファイルを読み、信頼できる提供元のものだけを入れてください。
Mac:Command+スペースで Terminal と入力 · Windows:下のコマンドは PowerShell では動きません。上の「AI にインストールしてもらう」を使ってください
repo を一時フォルダにダウンロードし、SKILL.md のあるフォルダだけをコピーします
コマンドを確認中 · 確認が通るとコピーボタンが出ます
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
自分でインストールした Skill は使えません。ディレクトリのプラグインに含まれる Skill だけが使えます
この Skill を ChatGPT で使うには、上の方法で ChatGPT デスクトップアプリを使ってください
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
この文をコピーして Codex のチャットに貼り付けます · ChatGPT デスクトップアプリ内の Codex、Codex CLI、IDE 拡張機能で使えます
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Codex は新しくインストールした Skill を自動で見つけます。表示されない場合は Codex を再起動してください · 使い方:ChatGPT では @ を入力して Skill を選びます · Codex CLI や IDE 拡張機能では /skills を実行するか、$ に続けて Skill 名を入力します
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
Skill は AI が従う指示です。入れる前にファイルを読み、信頼できる提供元のものだけを入れてください。
Mac:Command+スペースで Terminal と入力 · Windows:下のコマンドは PowerShell では動きません。上の「AI にインストールしてもらう」を使ってください
repo を一時フォルダにダウンロードし、SKILL.md のあるフォルダだけをコピーします
コマンドを確認中 · 確認が通るとコピーボタンが出ます
この文をコピーして Cursor のチャットに貼り付けます
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Cursor の Agent チャットで / を入力し kst-ai-assets-usage を選ぶ
Checked against official docs 2026年11月12日 · Source: cursor.com
Skill は AI が従う指示です。入れる前にファイルを読み、信頼できる提供元のものだけを入れてください。
Mac:Command+スペースで Terminal と入力 · Windows:下のコマンドは PowerShell では動きません。上の「AI にインストールしてもらう」を使ってください
repo を一時フォルダにダウンロードし、SKILL.md のあるフォルダだけをコピーします
コマンドを確認中 · 確認が通るとコピーボタンが出ます
Skill を読めるほかのアプリ向けです。この文には Claude Code、Codex、Cursor のパスが入っていて、AI が使っているアプリに合わせて選びます · Copy this sentence into the app’s chat
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Checked against official docs 2026年11月12日 · Source: code.claude.com, learn.chatgpt.com, cursor.com
Gemini CLI や GitHub Copilot など Agent Skills 対応の他のアプリはフォルダが違います。そのアプリのガイドを見て、SKILL.md のあるフォルダをコピーしてください。
2026年11月11日 に Claude Code、Codex、Cursor の公式ガイドで skill フォルダを確認
以下は提供元の repo の文章で、PopNx MCP の指示ではありません
最初の 8,000 文字を表示 · GitHub で全文を開く
---
name: kst-ai-assets-usage
description: Report which kasetto-installed skills and MCP servers are actually being used across the AI agents on this machine, and render a branded HTML dashboard of the result. Use whenever the user asks what agent assets they actually use, which skills or MCPs are dead weight, what to prune or clean up from kasetto.yaml, why their context is bloated with unused MCP servers, whether a skill has ever been invoked, or wants a usage report, audit, or dashboard of their agent setup. Also trigger on "kst ai assets usage", "ai asset usage", "what skills do I actually use", "which MCPs are worth keeping", "audit my agent assets", "am I using all these skills", or any request to review, trim, or justify an installed agent toolkit.
---
# kst-ai-assets-usage
Kasetto knows what you installed. Each agent knows what it ran. Nothing on the
machine joins the two, so installed-and-forgotten assets accumulate silently -
skills nobody has ever invoked, MCP servers loading tool definitions into every
request for a server the user stopped using months ago.
This skill closes that loop: read the lock, read each agent's local session
store, join them, and say plainly what is alive and what is dead.
## Run It
Two scripts, both standard-library Python 3, both offline.
```bash
python3 scripts/collect.py --out usage.json # scan + join
python3 scripts/render.py usage.json --out usage.html
```
`collect.py` is the slow one on a cold cache (a few seconds for a large history)
and near-instant afterwards - it caches per-file results in
`$XDG_CACHE_HOME/kasetto/usage-scan.json` keyed on size and mtime, so a rerun
only touches files the agent appended to. Delete that file to force a full
rescan.
Open the dashboard when it is written, then give the user the short version in
the terminal: how many assets are idle, the biggest surprises, and what you would
cut. The HTML is the artifact; your reading of it is the value.
## The Rule That Matters: Never Overstate Coverage
`collect.py` reports a `coverage` block listing every agent it knows about and
whether it found a readable session store. Honour it.
An asset the user drives daily from an agent that has no provider **looks
completely dead in this report**. If you present "55 skills never used" without
saying which agents that claim covers, and they act on it, they delete something
they rely on. That is the one way this skill can actively hurt someone.
So:
- State the coverage before any count that depends on it. The dashboard carries
this as a Review insight naming the unread agents rather than a banner, so it
is easy to scroll past - say it out loud in your summary
- Phrase the finding as "never invoked in *the agents kasetto can read*", not
"never invoked"
- When a `status: absent` agent is one the user actually uses, say so directly
and treat the idle list as a shortlist to review rather than a cut list
- A `verified_format: false` provider parsed a store whose layout has not been
confirmed against a real sample. A zero from it is weak evidence, not proof
## Reading the Output
Start with `insights` - `collect.py` computes the findings worth acting on
rather than leaving them to be read off a chart. Each carries a `severity`, a
`title` already phrased as a claim, a `detail` explaining why it matters, and
the `items` it refers to. Lead your summary with these, in order; they are
already sorted by severity. Then use the rest of the JSON to answer follow-ups.
Severity is a deliberately narrow vocabulary, and it drives colour in the
dashboard:
- `bad` (red) - Concrete waste, meaning a source repository with nothing in use. This
is the only thing red is spent on, because an unused skill is dormant, not
broken, and colouring it red would burn the strongest signal on the least
urgent finding
- `warn` (amber) - Act on it: idle MCP packs, coverage gaps, unverified formats
- `info` (cyan) - Context: concentration, freshly installed assets, unmanaged
finds
- `good` (green) - Nothing idle
Idle skills are grey throughout. Keep that distinction when you talk about them:
"dormant" and "worth a look", not "bad".
`usage.json` carries more than the dashboard shows. Worth reading directly when
the user asks something specific:
- `skills` / `mcps` - Per asset: `total`, `last_used`, `by_agent`, `days` (a
date-keyed histogram), plus `source` and `scope` from the lock. Skills carry
`age` and a `fresh` flag; MCP packs carry `by_server` and `by_tool`
- `by_source` - Per source repository, how many of its assets are live, fresh and idle.
**Usually the most useful view.** Assets go idle in clusters, because a repository
gets added for one skill and brings twelve. A wholly idle source is a single
`source:` block to delete rather than N rows to prune, so lead with this when
recommending cuts
- `activity` - Daily invocation totals for skills and MCPs, which is what
separates "used heavily last year" from "used steadily this week"
- `unmanaged` - Names called locally that are not in the lock: installed by hand
or shipped by the agent. A heavily used one is worth bringing under
`kasetto.yaml` so it syncs everywhere
- `counts` - Headline numbers, with `skills_idle` and `skills_fresh` already
separated
An MCP pack can merge several servers, so usage is counted per server and rolled
up. A pack alive on one server and idle on another is worth calling out; that
detail is invisible in the pack-level total.
## Recommending Removals
The idle list is a starting point, not a verdict. Before suggesting a cut:
- Skills cost almost nothing when idle - only their description is loaded until
invoked. An unused skill is clutter, not a real tax
- **MCP servers are different.** Their tool definitions load into context on
every single request whether or not you call them. An idle MCP pack is a
standing cost, so it is the far stronger cut candidate and worth leading with
- An asset installed last week that has not been used yet is not dead, it is new.
`collect.py` already separates these into `skills_fresh` using a 7-day
threshold, and the dashboard holds them in their own section. Never move one
into a cut list because its count is zero
- Some assets exist for rare high-stakes moments (an incident runbook, a release
procedure). Low count is the design, not a defect
To act on a decision, the existing CLI does it - `kst remove` edits
`kasetto.yaml` in place and preserves comments and key order, then `kst sync`
uninstalls the asset. Do not hand-edit the lock.
## Adding an Agent
The provider list is a table at the top of `scripts/collect.py`. Adding one is a
new entry plus, sometimes, a reader. `references/providers.md` documents each
known agent's store, which are verified against real data, and which are
inferred from documentation.
Two things to hold onto when you extend it:
**Match patterns, not schemas.** Session formats churn - Goose moved JSONL to
SQLite, OpenCode JSON to SQLite, Copilot CLI flat files to per-session
directories. What did *not* move through any of those migrations is the tool-name
string. Matching `mcp__<server>__<tool>` survives a schema rewrite that would
break a field-by-field parser. Resist the urge to "properly" parse these files.
**Anchor on an invocation position.** Session logs also record the tools
*offered* each turn, as arrays of names repeated on every request. An early
version matched the bare `mcp__x__y` string anywhere and reported 38,743 calls
for a server with zero real invocations. Every pattern must require a key
position (`"name": "mcp__..."`), and any new provider needs a sanity check
against an independent count before its numbers are trusted.
## Scope
Skills and MCP servers only. The other four asset kinds are deliberately out:
- Commands - In the logs they are mostly agent built-ins rather than kasetto
assets, so a count would mostly measure the agent
- Instructions - Injected as context and never "invoked", so usage質問、コツの共有、使えないことの報告ができます。「役に立った」の多い順です。