Price Monitor

Monitor website prices, inventory, and content changes using browser automation. Use when tracking e-commerce prices, competitor monitoring, stock alerts, or...

Security Scan

The skill's files mostly match the stated purpose (price monitoring), but there are clear inconsistencies (undeclared dependency on an external 'agent-browser' binary and SKILL.md promises alert integrations that the shipped script does not implement) that should be resolved before installing.

Purpose & Capability

The name/description align with the included Python script: it reads a CSV of product URLs, uses browser automation (via an external agent-browser binary) to extract prices, logs history, and reports changes. However, SKILL.md advertises email/Discord webhook alerts and several alert options that are not implemented in scripts/monitor_prices.py — the script only logs to CSV and prints alerts to console. Also SKILL.md recommends proxies/UA rotation, but the code does not implement these features.

Instruction Scope

Runtime instructions focus on visiting product pages, taking snapshots, extracting text, and running the Python monitor. The code does exactly that (calls agent-browser, parses output, writes price-history.csv). There is no code that reads unrelated system files or transmits data to arbitrary endpoints. Note: if you point the skill at pages behind authentication or internal URLs, the skill will capture page content you give it access to — expected but worth noting.

Install Mechanism

This is instruction-only with a local Python script (no install spec). However, both SKILL.md and the script depend on an external 'agent-browser' CLI being present and callable via subprocess. The skill metadata lists no required binaries — that is an inconsistency. The absence of a declared dependency on agent-browser is a risk: the script will fail or behave unexpectedly if agent-browser is not present or is a different program, and it places implicit trust in whatever 'agent-browser' is installed on PATH.

Credentials

The skill declares no required environment variables or credentials and the Python code does not read env vars. SKILL.md mentions SMTP and Discord webhooks for alerts (and suggests proxies), but those integrations are not implemented and no credentials are requested. This mismatch could confuse users and may lead to ad-hoc modifications that introduce credential handling without clear guidance.

Persistence & Privilege

The skill is not always-enabled and does not request elevated or persistent system privileges. It writes price-history.csv to the working directory (normal for this use). It does not modify other skills' configs or system-wide settings.

What to consider before installing

This skill appears to generally do what it claims (visit pages and log prices), but you should resolve a few issues before trusting it: - Ensure a legitimate 'agent-browser' CLI is installed and review what that binary does. The Python script calls subprocess.run(['agent-browser', ...]) but the skill metadata did not declare agent-browser as a required binary. If a malicious binary named 'agent-browser' were on PATH it could be invoked. - Be aware the shipped script only logs to CSV and prints alerts — SKILL.md mentions email/Discord/webhooks and proxy/UA rotation, but those features are not implemented. If you need them, implement them yourself and avoid storing credentials in plaintext or hard-coding webhooks. - Confirm the working directory where the script runs. It will write price-history.csv and will read whatever products CSV you pass; do not point it at sensitive internal URLs unless you intend it to capture those pages. - Rate limiting: the script sleeps 2 seconds between checks but SKILL.md recommends slower polling (30s+). Adjust sleep/cron frequency to avoid being blocked by sites and to respect ToS. If you proceed, audit and control the agent-browser binary, run the script in a restricted environment, and add explicit, secure handling for any alert credentials you add. If you want a cleaner install, insist the skill metadata declare required binaries (agent-browser) and any needed env vars for alerts.

Overview

Automated price and content monitoring skill using agent-browser. Tracks price changes, stock availability, and content updates on any website with configurable alerts and history logging.

Quick Start

# Monitor a product price
agent-browser open "https://example.com/product/123"
agent-browser snapshot -i
agent-browser get text @e1  # Get price element

Core Workflows

1. Single Product Price Check

Use case: Check current price of a specific product

# Navigate to product page
agent-browser open "<product-url>"

# Get page snapshot to find price element
agent-browser snapshot -i

# Extract price (use appropriate ref from snapshot)
agent-browser get text @e1

# Optional: Save to history log
echo "$(date), $(price)" >> price-history.csv

2. Multi-Product Monitoring

Use case: Track prices across multiple products/competitors

Create a products.csv with URLs and price selectors:

url,selector,name
https://site-a.com/product1,.price-tag,Product A
https://site-b.com/item2,#price,Product B

Run monitoring script:

python scripts/monitor_prices.py products.csv

3. Stock/Inventory Alerts

Use case: Get notified when out-of-stock items become available

agent-browser open "<product-url>"
agent-browser snapshot -i
agent-browser get text @e1  # Check for "In Stock" or "Out of Stock"

4. Price History Tracking

Use case: Build historical price data for analysis

Script automatically logs:

• Timestamp
• Product name/URL
• Current price
• Stock status

Output: price-history.csv or JSON format

Scripts

monitor_prices.py

Main monitoring script that:

• Reads product list from CSV
• Navigates to each URL
• Extracts price using CSS selector
• Logs results with timestamp
• Detects price changes
• Optional: Send alerts on significant changes

Usage:

python scripts/monitor_prices.py products.csv [--alert-threshold 10]

Arguments:

• products.csv - Product list with URLs and selectors
• --alert-threshold - Percentage change to trigger alert (default: 10%)

Configuration

Product List Format (CSV)

url,selector,name,min_price,max_price
https://amazon.com/dp/B08N5WRWNW,.a-price-whole,Sony Headphones,50,150
https://bestbuy.com/site/12345,.priceView-hero-price,TV,200,500

Alert Options

• Email alerts - Configure SMTP settings
• Discord webhook - Post to Discord channel
• File logging - Append to CSV/JSON
• Console output - Print changes to terminal

Best Practices

1. Rate limiting - Add delays between requests (30s+ recommended)
2. Error handling - Handle page load failures gracefully
3. Selector stability - Use stable CSS selectors, avoid dynamic classes
4. Headless mode - Run browser in headless mode for automation
5. Schedule wisely - Check prices during business hours for accuracy

Example: Daily Price Check Cron

# Run every day at 9 AM
0 9 * * * cd /path/to/skill && python scripts/monitor_prices.py products.csv

Troubleshooting

• Element not found: Re-run snapshot to get updated refs
• Price format issues: Adjust selector or parse with regex
• Page load timeout: Increase timeout or add wait condition
• Blocked by site: Add delays, rotate user agents, or use residential proxy