Ticket Signal Watch

This skill is for reliable ticket notifications, not automated checkout.

Use it when:

• a user wants to monitor ticket sale or restock signals
• a backend already has ticket data and needs a push-ready event format
• you want OpenClaw to check official pages, search pages, or other text sources for ticket signals

Files

• Skill root: {baseDir}
• Script: {baseDir}/scripts/watch_ticket_pages.py
• Target example: {baseDir}/config/targets.example.json
• Notifier example: {baseDir}/config/notifiers.example.json
• Default state path: {baseDir}/state/state.json

Operating model

Treat the workflow as three layers:

1. collector
   • fetch page text or consume backend data
2. signal engine
   • decide whether the change is meaningful
3. notifier
   • push the resulting event to OpenClaw channels, webhooks, or another downstream system

The ideal production setup is:

• use backend data if available
• fall back to page checks only when needed
• keep collection and notification decoupled

Recommended usage

Run the script with a target config and a writable state file:

python3 "{baseDir}/scripts/watch_ticket_pages.py" \
  --config "{baseDir}/config/targets.example.json" \
  --state "{baseDir}/state/state.json" \
  --json

If alerts is empty, do not send a notification.

If alerts is non-empty, forward the structured result to:

• an OpenClaw channel
• a webhook
• a file/queue processor

Output expectations

The script should produce:

• results: per-target check results
• alerts: only meaningful changes worth notifying
• summary: short human-readable summary

Each alert should contain enough information to route downstream:

• name
• platform
• url
• signal_hits
• signal_level
• alert_reasons

Configuration guidance

Prefer official detail pages over generic search pages.

Use:

• require_all for identity words that must be present
• match_any for actionable signal words
• signal_keywords.high for the strongest signals

Examples of strong signals:

• 立即购买
• 立即预订
• 可选座
• 有票
• 开售
• 预售中
• 补票
• 回流
• 加场

Guardrails

• Do not treat generic page changes as sale signals.
• Search pages are weaker than detail pages.
• Add cooldown, dedupe, and jitter before high-frequency polling.
• If a platform starts returning anti-bot pages or challenge pages, mark that explicitly instead of claiming success.
• This skill is for notification workflows; do not imply that it can safely complete checkout automatically.