Scrapling Yoo

Security checks across malware telemetry and agentic risk

Overview

This is a real scraping skill, but it normalizes anti-bot bypass, Cloudflare solving, proxy rotation, broad crawling, and local writes without enough per-use safeguards.

Install only if you need advanced scraping and have authorization for the target sites. Prefer normal fetching first, set explicit domains, depth, concurrency, and delays, and use stealth, Cloudflare solving, proxy rotation, login sessions, or protected-page scraping only with clear permission. Use an isolated environment and clean up scrape outputs, downloaded media, cookies, and crawl checkpoints when finished.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Context-Inappropriate Capability

Medium

Confidence: 87% confidence
Finding: The media downloader recipe goes beyond scraping guidance by showing how to fetch remote content and write it directly to the local filesystem. In a guidance-only skill, this expands capability into persistence of untrusted data, which can enable disk abuse, storage of malicious content, or unsafe downstream handling if copied into agent workflows without safeguards.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The checkpoint reset example includes destructive local deletion via shutil.rmtree without safety checks or warnings. Even though the example targets a fixed path, patterns like this are easily reused in broader agent contexts and normalize unsafe deletion behavior unrelated to the stated scraping-guidance purpose.

Vague Triggers

Medium

Confidence: 83% confidence
Finding: The description is broad enough that an agent could select this skill for many loosely related web tasks, including sensitive or higher-risk scraping scenarios. Overbroad triggering increases the chance of unnecessary activation of powerful scraping and anti-bot functionality in contexts where it should not be used.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The skill gives concrete guidance for stealth scraping, TLS fingerprint spoofing, proxy rotation, and automatic Cloudflare/Turnstile solving, but does not place prominent warnings and hard constraints directly alongside those instructions. In context, this materially lowers the barrier to bypassing site protections and can facilitate unauthorized access, evasion of rate limits, and violations of law, terms of service, or account controls.

Missing User Warnings

Medium

Confidence: 83% confidence
Finding: This documentation explicitly exposes network-fetching, crawling, and anti-bot/Cloudflare-bypass capabilities without any safety guidance about authorization, terms-of-service compliance, privacy, or data transmission. In an agent skill, that omission materially increases the chance of misuse for unauthorized scraping or collection of sensitive data, especially because the examples normalize stealthy access to protected sites.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The markdown presents recursive file deletion with no cautionary note, confirmation step, or scoping guidance. Documentation is often copied verbatim into automation, so omitting warnings materially increases the chance of accidental destructive actions by users or agents.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal