marketplace-monitor
Skill0 starsChanged 4 months agoMITView on GitHub
Stand up a recurring watcher for any online marketplace search that alerts on new matching listings and drafts a first message to the seller. Use when the user wants to watch a marketplace (Facebook Marketplace, Craigslist, Kijiji, eBay, AutoTrader) for items matching criteria, or says "tell me when something like X gets listed".
In this repository6 other files it gives its agents.
- book-to-skillSkill
- buyback-time-auditSkill
- market-pulseSkill
- meeting-intelSkill
- relationship-radarSkill
- weekly-briefSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 17
Not a chat answer - a small, working monitor:
- Distinctive: line 35
Check authentication reality before writing code: Facebook Marketplace requires a logged-in session (persist a browser profile the user …
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Marketplace Monitor
Turn "keep checking the marketplace for me" into a script that actually keeps checking.
Distilled from a Playwright monitor that checked Facebook Marketplace every 30 minutes for free kids' sports gear and messaged each new poster. The pattern generalizes to any marketplace with a search URL.
What this skill builds
Not a chat answer - a small, working monitor:
- A scraper script (Playwright for JS-heavy sites, requests + parsing where HTML suffices) that runs one search and extracts listings: title, price, location, URL, posted date.
- A seen-listings store (SQLite) so each run only surfaces new items.
- A notifier for new matches, using what the user has: Telegram bot, email, macOS notification, or append to a watched file. Ask which.
- A scheduler: launchd plist on macOS or cron elsewhere, default every 30 minutes, respecting quiet hours (default 22:00-07:00).
- Optionally, a drafted first message per new listing, saved next to the alert. Off by default.
Workflow
- Get the target: marketplace, search terms, price cap, location radius, any dealbreakers. Build the canonical search URL and confirm it with the user by opening it once.
- Check authentication reality before writing code: Facebook Marketplace requires a logged-in session (persist a browser profile the user logs into once); Craigslist and Kijiji generally do not. Say which path applies. If the site blocks automation entirely, say so and offer the closest alternative (RSS, saved-search emails, a different market).
- Build the scraper, run it once live, and show the user the first results before installing any schedule. A monitor that has never succeeded once must not be scheduled.
- Install the schedule and tell the user exactly how to stop it (the one-line unload/removal command).
- Auto-messaging sellers: only wire this if the user explicitly asks, and always show the message template for approval first. Default is draft-only.
Rate limits and conduct
- Minimum interval 15 minutes. One search page per run, no pagination crawls. The monitor should look like a person checking, not a bot harvesting.
- Respect the site's terms; when the user asks for volume or scraping breadth that crosses into abuse, decline and say why.
Output
A working directory containing the scraper, the store, the notifier config, the schedule file, and a README with: what it watches, how often, how to stop it, where alerts go.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

