/competitor-monitor

name: competitor-monitor
description: Monitor competitor websites for changes. Visit a list of URLs, extract pricing, features, positioning, and key content, compare against previous snapshots stored locally, and generate a change report summarizing what's different. Use when the user says "check competitors", "what changed on their site", "monitor these URLs", or wants periodic competitive intelligence. Requires the hanzi browser automation MCP server and Chrome extension.
category: marketing

Competitor Monitor

You monitor competitor websites and report what changed. You visit each URL, extract the important content (pricing, features, positioning, messaging), compare it against the last saved snapshot, and produce a clear change report.

Tool Selection Rule

• **Prefer existing tools first**: If a page is public and simple, try `WebFetch` or `curl` before opening a browser. Use Hanzi only when the page requires JavaScript rendering, authentication, or interactive elements (tabs, accordions, lazy-loaded sections).
• **Use filesystem tools** to read/write snapshots — never store snapshots in the browser.
• **If a site blocks or shows a CAPTCHA**, stop that URL and move to the next. Report the failure.

Before Starting — Preflight Check

Try calling `browser_status` to verify the browser extension is reachable. If the tool doesn't exist or returns an error:

> **Hanzi isn't set up yet.** This skill needs the hanzi browser extension running in Chrome. > > 1. Install from the Chrome Web Store: https://chromewebstore.google.com/detail/hanzi-browse/iklpkemlmbhemkiojndpbhoakgikpmcd > 2. The extension will walk you through setup (~1 minute) > 3. Then come back and run this again

---

What You Need From the User

1. **URLs** — list of competitor pages to monitor (pricing pages, feature pages, landing pages, etc.) 2. **Focus areas** (optional) — what to pay attention to: pricing, features, positioning, team size, integrations, messaging, or "everything" 3. **Label** (optional) — a name for this monitoring set (e.g., "competitor-pricing", "market-landscape"). Defaults to "default".

Optional:

• Specific sections or elements to watch (e.g., "only the pricing table", "the hero section tagline")
• Whether to take screenshots for visual comparison
• Authentication details if any pages require login

---

Phase 1: Prepare the Monitoring Run

1a. Load Previous Snapshots

Snapshots are stored in `~/.hanzi-browse/competitor-monitor/{label}/`. Each URL gets a file named by its sanitized hostname + path.

mkdir -p ~/.hanzi-browse/competitor-monitor/{label}
ls ~/.hanzi-browse/competitor-monitor/{label}/ 2>/dev/null || echo "NO_PREVIOUS_SNAPSHOTS"

For each URL, check if a previous snapshot exists:

cat ~/.hanzi-browse/competitor-monitor/{label}/{sanitized_filename}.json 2>/dev/null || echo "NO_SNAPSHOT"

If no previous snapshots exist, this is a **baseline run** — you'll capture the initial state without generating a diff.

1b. Plan the Extraction

For each URL, determine what to extract based on the page type and user's focus areas:

| Page Type | What to Extract | |-----------|----------------| | **Pricing page** | Plan names, prices, billing periods, feature lists per tier, CTAs, free tier details, enterprise contact options | | **Features page** | Feature names, descriptions, categories, "new" or "coming soon" badges, comparison tables | | **Landing/home page** | Hero headline, subheadline, value propositions, social proof (logos, testimonials, stats), CTAs | | **About/team page** | Team size, key hires, office locations, funding mentions | | **Blog/changelog** | Latest 3-5 post titles, dates, and summaries | | **Integrations page** | List of integrations, categories, "new" badges | | **Docs/API page** | Navigation structure, new sections, deprecation notices |

Present the plan: "I'll visit these N URLs and extract [focus areas]. Previous snapshots: [found/not found]. Ready to proceed?"

**Wait for user confirmation before visiting any URLs.**

---

Phase 2: Visit and Extract (browser via Hanzi)

Visit each URL using `browser_start`. Run up to 3 URLs **in parallel** — each gets its own browser window.

For each URL:

browser_start({
  task: "Visit this page and extract all [focus areas]. Read the full page content including any sections behind tabs, accordions, or 'show more' buttons. Return structured data with: page_title, extraction_date, and each content section with its heading and text.",
  url: "{competitor_url}",
  context: "Focus areas: {focus_areas}. Extract exact text — do not paraphrase. Include prices with currency symbols. Expand any collapsed sections."
})

After `browser_start` returns: 1. Parse the result to extract structured content 2. If the user requested screenshots, call `browser_screenshot` for each page 3. Call `browser_stop` with `remove: true` to clean up

Extraction Format

Structure the extracted data consistently:

{
  "url": "https://competitor.com/pricing",
  "extracted_at": "2026-04-02T10:30:00Z",
  "page_title": "Pricing - Competitor",
  "sections": [
    {
      "name": "Plans",
      "content": [
        {
          "plan": "Starter",
          "price": "$9/mo",
          "billing": "billed annually",
          "features": ["Feature A", "Feature B", "5 users"]
        }
      ]
    },
    {
      "name": "Hero",
      "headline": "The fastest way to do X",
      "subheadline": "Used by 10,000+ teams"
    }
  ]
}

Error Handling

• **Page blocked / CAPTCHA**: Skip, note in report, move to next URL
• **Page not found (404)**: Record as "page removed" — this itself is a significant change
• **Timeout**: Call `browser_screenshot` to capture current state, then `browser_stop`. Retry once. If it fails again, skip.
• **Login required**: Stop and ask the user for credentials. Pass via `context` field, never in `task`.

---

Phase 3: Compare Against Previous Snapshots (no browser)

For ea