REA
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
SEO foundation checks as an Agent Skill: give it a URL, get a crawler's-eye pass/fail report with evidence. 26 structural checks, zero LLM, deterministic.
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/tigerless-labs-seo-ops)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 seo-ops | rea | andrej-karpathy-skills | graphify | MoneyPrinterTurbo | |
|---|---|---|---|---|---|
| 作者 | tigerless-labs | morluto | multica-ai | Graphify-Labs | harry0703 |
| Skill 数 | 1 | 1 | 1 | 1 | 1 |
| GitHub スター | 989 | 463,670 | 233,682 | 148,887 | 143,987 |
| 最終コード更新 | 2026年11月6日 | 2026年11月17日 | 2026年4月20日 | 2026年11月17日 | 2026年11月16日 |
| ライセンス | – | MIT | – | Apache-2.0 | MIT |
tigerless-labs 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 CodeCodex
提供元の README と repo のフォルダから判断
GitHub account tigerless-labs (organization) is only 5 months old: take extra care
★ 989 stars · code last changed 2026年11月6日
2 scripts that can run on a computer: scripts/config.py, scripts/run.py
Scripts that go online: scripts/run.py
Scripts that touch settings folders (~/.claude or settings.json): scripts/run.py
Checked automatically from the repo files at commit 777bde4 (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 seo-ops folder again. The ZIP must have a seo-ops folder with SKILL.md at the top level, no loose files
+Create skillUpload a skillpick seo-ops.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、/seo-ops 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。
コマンドを確認中 · 確認が通るとコピーボタンが出ます
提供元の README から引用。提供元のコードをダウンロードして実行します
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
seo-ops フォルダを ~/.agents/skills/seo-ops 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 seo-ops 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: seo-ops
description: Structural checklist and checker script for SEO foundation engineering. Use when reviewing or accepting a site/page for SEO/GEO structural compliance (robots, sitemap, canonical, 301 canonicalization, JSON-LD, OG, hreflang, CWV, AI crawler access, rendering strategy), when running the machine checks to produce a report, or when judging whether a frontend change will affect indexing and citation. Also for answering "what exactly does this check require" and "why does it exist".
compatibility: Requires Python 3.9+, requests, PyYAML, and network access to the target site
---
# seo-ops · SEO foundation engineering
Review a site's templates and output against the **C set** (30 structural checks), or run the checker to produce an acceptance report.
## How to work
When this skill triggers, introduce it in **one sentence** — "30 structural checks (the C set): can crawlers
fetch, read, index, and cite this site" — then immediately ask **one question** (as selectable options if your
environment has an option/question UI):
> **What should I check?**
> 1. **A URL** — live domain or local deployment (`http://localhost:<port>`), both fine; produces a report
> 2. **The code** — review templates/pages in this repo against the checklist, no script
That one answer decides the path; don't ask anything else before starting — **no config is needed**, the checker
runs with defaults on a bare `--target`. Never bring configuration up on your own; it only becomes relevant if
the user asks for the extras (CrUX field data, IndexNow, multi-site roster — see "**Configuration**").
| The user gave | Path | Section to follow |
|---|---|---|
| a domain / URL (live or localhost) | run the checker, produce a report | "**Run the checker**" → then "**Read the report**" |
| code, templates, a PR, page files | no script; review the machine-readable output surface against the C set item by item | "**Scope**" sets the boundary; `references/checklist/checklist.md` lists the items — but judge each one by its canonical doc `references/checklist/references/C<N>.md`, not by the index row alone |
Once intent is confirmed, **do not keep asking for permission**. Everything else:
the user asks why a check exists / how to fix it → "**After the report**";
dependencies won't install → "**Environment**".
Both paths **treat [references/checklist/checklist.md](references/checklist/checklist.md) as authoritative** —
never judge from memory; if unsure about an item, read `references/checklist/references/C<N>.md`.
Cite items by ID (`C12`); IDs are permanent — never renumber or reorder.
### After the report
**The report itself is the conclusion — don't walk through it item by item unprompted.** Each row carries a **Docs** column: the GitHub link to that check's canonical doc
— that link is for **the person reading the report** (it opens wherever the report is sent). **What you yourself read is the local copy**:
`references/checklist/references/C<N>.md`.
**When the user asks about a check, read `references/checklist/references/C<N>.md` before answering**
(for C12, read `references/checklist/references/C12.md`): "why it's a problem" comes from `## Introduction`,
"how to fix it" from `## Implementation guidance`. **Never answer by inferring from the evidence** — evidence only says which pages failed;
it can't tell you why the check exists, what the correct fix is, or where the authoritative reference lives.
### Environment
```bash
python3 -c "import requests, yaml" || pip install -r scripts/requirements.txt
```
On `externally-managed-environment` (PEP 668, common on Debian/Ubuntu), use the system packages
`apt install python3-requests python3-yaml` or create a venv — **do not add `--break-system-packages`**;
that gambles the system Python for a two-line dependency.
## Scope
**SEO foundation engineering** = the machine-readable structure a site must have so that search engines and AI retrieval can **fetch, read, index, and cite** it.
**What it checks**: the machine-readable surface of the final HTTP/HTML output — regardless of who produced it (backend templates, frontend code,
third-party scripts); if a crawler can see it, it's covered. The boundary is "machine-readable surface / human-perceivable surface", not frontend/backend.
**What it doesn't check**: styling, interaction, UX, code quality; content truthfulness and quality; ranking and traffic; process.
**All green ≠ fully compliant.**
**The human gate is a separate set**: the R1-R8 prohibitions (buying links, cloaking, fake structured data, PII in prompts, etc.) live in
[redlines.md](redlines.md); the machine doesn't check them. Read them before touching content strategy, links, or structured data.
## What's here
| Path | What it is |
|---|---|
| [references/checklist/checklist.md](references/checklist/checklist.md) | The **C set** index: 30 checks, split into site-level / per indexed page / conditional |
| `references/checklist/references/C<N>.md` | Each C check's canonical doc: `## Introduction` (why) + `## Implementation guidance` (how to fix) + authoritative references |
| `references/content/` | The **T set**: for tracing a red C back to "who owes the input"; outside this skill's duties |
| `scripts/run.py` · `scripts/config.py` | Check logic / check parameters |
| `references/ai-crawlers.yaml` | The AI crawler UA roster for C1, read at runtime |
| `references/{sites,config}.example.yaml` · `references/.env.example` | Three config templates, see "**Configuration**" |
`<N>` maps one-to-one to the IDs in the checklist; C12's canonical doc is `references/checklist/references/C12.md`.
Gaps in the numbering are normal (retired IDs are never recycled).
**The skill directory is read-only — zero writes.** Updates overwrite the whole package; anything written inside will be lost.
## Run the checker
**Confirm the target is an origin → run.** No config file is required and none should be asked about —
defaults cover everything; the optional files exist for users who come asking for more (see "Configuration").
**One "site" = one origin** (scheme + host[:port], no path/query). Subdomains count as separate sites
(`blog.` / `docs.` are each one); the apex domain and www don't count as two — they should 301 to a single canonical host,
which is exactly what C3 checks, so only the one you pick counts.
### Single site: no config file needed
```bash
python3 scripts/run.py --target https://example.com # live site
python3 scripts/run.py --target http://localhost:3000 # local deployment, auto-detected as local mode
```
`--target` must be an origin; a URL with path/query exits with an error immediately — no guessing.
**Public staging domains are not supported** — they'd be tested as production domains, and the canonicalization check (C3) would false-red for sure.
To test before launch, pick one: run a local deployment at `http://localhost:<port>`, or wait and run against the production domain after launch.
### Multi-site: run a batch in one go
"Multi-site" = you have several domains to accept at once (e.g. `example.com` + `example.org` + `blog.example.com`),
don't want to retype `--target` every time, and want each site to carry its own rendering strategy and must-test pages. After creating `sites.yaml` (see "Configuration"):
```bash
python3 scripts/run.py # run every site in the roster
python3 scripts/run.py --site <id> # run just one
```
Each entry takes: `id` (used for the report filename and the `checks` table's site column), `production` (origin, required),
`rendering` (ssr/ssg/isr; C15 branches on it), `sitemap` (defaults to `<production>/sitemap.xml`),
`samples` (must-test pages fetched in addition to the sitemap; mark `ymyl: true` to trigger the C21 human review).
If the roster can't be found, the script exits and prints where to create it — it never silently runs empty.
### Common flags
One-off overrides: `--page-sample N` `--sitemap-sample N` `--max-pages N` `--sleep S` `--w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。