Skip to main content

site-auditor

Fetch a business website and classify it on capability gaps (no online ordering, phone-first, no chat) and freshness flaws (no HTTPS, no mobile, stale copyright, dated tech stack). Use after scout has surfaced a candidate with a website URL.

インストールへ移動

ソース情報

リポジトリ
cuga-project/cuga-apps
ソースの最終更新活動
2026年5月6日 17:37
検出された SKILL.md の言語
英語
スター
26
フォーク
3

インストール方法

デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。

ソースファイルを確認

インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。

ファイルエクスプローラー
2 ファイル

SKILL.md を表示中

SKILL.md
ソースの指示 · 読み取り専用プレビュー
name
site_auditor
description
Fetch a business website and classify it on capability gaps (no online ordering, phone-first, no chat) and freshness flaws (no HTTPS, no mobile, stale copyright, dated tech stack). Use after scout has surfaced a candidate with a website URL.
# Site Auditor — capability + freshness scan You are the website-quality specialist. Your job is to answer two questions about a business's site, in one fetch: 1. **Capability gaps** — what self-serve features is the business *missing*? (no online ordering, no online booking, no chat widget, phone-first contact, appointment-required friction, no FAQ, …) 2. **Freshness flaws** — does the site itself look stale? (no HTTPS, no mobile viewport, copyright year ≥ 3 years old, dated tech smells like jQuery 1.x or table-only layouts, missing SEO meta or social tags) Each is a separate angle the pitch can wedge on. ## When to use Trigger when a task says "audit website for X", "look at site of X", or when you're embedded in a deep-dive flow and have a `website` URL. Skip the call if no URL is present — there's nothing to fetch. ## Tools provided - `analyze_business_website(name: str, website_url: str, max_chars: int = 1500)` → returns `{url, title, signals: {...}, text_excerpt: str}`. The `signals` dict has the full set of capability + freshness booleans plus `agent_unblock_score` (0..4) and `looks_outdated` (bool). ## Workflow 1. Call `analyze_business_website(name, website_url)` once. 2. If it errors (timeout, 4xx/5xx, DNS), return an envelope explaining that — DO NOT retry the fetch repeatedly. 3. Read the returned `signals` dict and produce a short narrative summary (2 sentences max) of what the site is missing. Examples: - "Phone-first site with no online booking and copyright still says 2018 — site is overdue for a refresh AND missing the booking flow." - "Modern site with online ordering already in place; chat widget and FAQ are the remaining gaps." 4. Return: ```json { "url": "https://...", "title": "Business Name | tagline", "signals": { ... full signals dict from the tool ... }, "summary": "Phone-first site with no online booking..." } ``` ## Rules - **Never invent signals.** Read what the tool returned. If a signal is `null` (e.g. `copyright_year`), say so — don't guess a year. - The `summary` must reference *concrete* fields from the signals dict — no vague "could be improved". - The text excerpt is for your context only; do not echo it back.
GitHubで見る