askill
bargain-finder

bargain-finderSafety 90Repository

Find local marketplace bargains via browser tools and send top deals to Telegram

8 stars
1.2k downloads
Updated 2/8/2026

Package Files

Loading files...
SKILL.md

Overview

Use this skill to find the best local bargains on Facebook Marketplace and Craigslist with the browser tools only. The goal is a short list of top 3–5 deals with a mini-report per deal explaining why it is a great bargain.

Default search center: 957 Hayes St, San Francisco, CA (94117).

Prerequisites

  • Chrome running with the OpenCode browser extension (see browser skill).
  • Telegram configured (see telegram skill).

Inputs (from the user prompt)

  • Items: list of items to search for (required).
  • Optional filters: brand preferences, max price caps, condition notes.
  • Excluded keywords (default): broken, for parts, repair, not working.

Bargain Hallmarks (stupid simple)

  1. Distance first
    • 0–1 mile: always best
    • 1–3 miles: acceptable
    • 3–5 miles: only if insanely cheap
  2. Price vs typical
    • Glance at the first 8–10 results to estimate a typical price.
    • If distance is 3–5 miles, only keep listings priced ~50% or less of typical.
  3. Clean condition
    • Clear photos and description, no damage language.
  4. Urgency signals
    • “moving,” “must go,” “today,” “pickup only,” “OBO.”

Workflow (browser tools only)

1) Build search URLs

For each item in the prompt:

  • Facebook Marketplace: https://www.facebook.com/marketplace/sanfrancisco/search?query=<item>
  • Craigslist: https://sfbay.craigslist.org/search/sss?query=<item>&search_distance=5&postal=94117

2) Apply filters

  • Set radius to 5 miles and sort by closest (if available).
  • If the UI resets location, set it back to 957 Hayes St.

3) Collect candidate listings

For each item, scan the first page and open promising listings in new tabs. Capture:

  • title
  • price
  • distance
  • condition notes
  • link

Exclude listings that contain the default broken keywords.

4) Rank and pick top 3–5

  • Rank by distance bucket first, then price vs typical.
  • Keep only “insanely cheap” deals if they are 3–5 miles away.
  • Add urgency/quality signals as tie-breakers.

5) Send Telegram summary

Send a concise report with 2–4 lines per item, capturing the price, distance, condition signals, and why it is a great bargain.

Telegram Output Format

Use Markdown:

*Daily Bargains (SF 957 Hayes St)*

1) $120 — Yamaha receiver + speakers (0.8 mi) [FB link]
   Why it’s a deal: ~50% below typical; clean photos, complete set.
   Condition: described as “like new,” no damage language.
   Notes: pickup today, seller says “must go.”

2) $80 — Solid wood coffee table (2.2 mi) [CL link]
   Why it’s a deal: price below similar listings nearby.
   Condition: solid wood, minor scuffs only.
   Notes: moving sale, quick pickup.

Daily Scheduler Prompt Template

---
schedule: "0 9 * * *"
---

@bargain-finder
Find top 3–5 deals near 957 Hayes St, San Francisco. Use browser tools only.
Search Facebook Marketplace and Craigslist for: <items list>.
Apply distance rule (1–3–5 miles), exclude broken listings, and send a short report per deal with why it is great.
Send results via Telegram.

Common Gotchas

  • Facebook Marketplace requires an authenticated session in Chrome.
  • Filters sometimes reset when switching queries; re-check radius.
  • If a listing lacks distance, skip it unless price is obviously insane.

When Done

Always release the browser lock:

Close the tab
browser_browser_release()

Install

Download ZIP
Requires askill CLI v1.0+

AI Quality Score

88/100Analyzed 3/27/2026

Well-structured skill with clear workflow for finding marketplace bargains. Provides specific URLs, step-by-step instructions, Telegram output format, and a scheduler template. The location-specific default (957 Hayes St, SF) and .opencode path suggest some internal customization, but the approach is systematic and reusable. Tags don't match the skill content well (ci-cd/github-actions for a browsing skill), but the core content is practical and actionable.

90
95
85
90
95

Metadata

Licenseunknown
Version-
Updated2/8/2026
Publisherbenjaminshafii

Tags

ci-cdgithub-actionsprompting