جاك العلم
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only shopping assistant is coherent and purpose-aligned, but users should notice that it browses external sites, shares preferences across sub-agents, and creates persuasive buy-link reports.
This looks reasonable for shopping research. Do not provide payment, bank app, login, or highly personal details; treat the report as advice rather than a purchase authorization; and verify seller, price, warranty, delivery, and return terms before clicking a buy link.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Your product query, budget, and preferences may be sent to search engines or retailer pages during research.
The skill explicitly uses web search, page fetching, and browser automation to research shopping options across retailer and search sites.
TOOLS: sessions_spawn (sub-agents), web_fetch, web_search, camofox_* ... Store Access Methods ... All others | web_fetch first, camofox fallback
Use it for normal shopping research, but avoid entering private account, payment, or banking details and verify the final retailer page yourself before buying.
Shopping preferences, budget, and use case may be shared among spawned sub-agents, and mistakes in one agent's findings could influence the final report.
The workflow intentionally passes the user's shopping request through multiple research roles before producing a final recommendation.
The main assistant acts as Router/Orchestrator, spawning sub-agents to research products or services
Keep sensitive personal information out of the shopping request and review the cited sources in the final report.
Opening the report can contact external servers, which may reveal basic browser/IP/referrer information.
The saved HTML report may load Google Fonts and retailer-hosted product images when opened.
<link href="https://fonts.googleapis.com/css2?family=Rubik..." ... Product images: Use product images from retailer sites if found.
If privacy or offline viewing matters, use local fonts/images or remove external resources from the generated report.
The presentation may make a recommendation feel more certain than the underlying web research warrants.
The report is designed to make purchasing easy and persuasive, with direct buy buttons and ranked recommendations.
One-tap buy: The purchase button is large, colored, and goes directly to the product page.
Before purchasing, independently confirm price, seller reputation, warranty, delivery time, return policy, and whether the product still fits your needs.