ClawHub
Back to skill

Security audit

Meta Ads Collector Adarsh

Security checks across malware telemetry and agentic risk

Overview

This is a read-only Meta Ad Library collector with expected external API use and no evidence of hidden, destructive, or deceptive behavior.

Install only if you are comfortable using a Meta access token for Ad Library queries. Use a least-privilege Meta app/token, do not put secrets in prompts or logs, and verify whether US-only results are acceptable for your audit before relying on the output.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Missing User Warnings

Low
Confidence
89% confidence
Finding
The skill documentation states that brand and optional domain inputs are sent to Meta's external Ad Library API and that Meta credentials are required, but it does not clearly warn users/operators about this outbound data transmission or dependency on third-party credentials. This is primarily a transparency and privacy-risk issue: operators may unintentionally send sensitive client identifiers or internal targets to an external service without explicit notice.

Natural-Language Policy Violations

Medium
Confidence
82% confidence
Finding
The query parameter hard-codes ad_reached_countries to US-only without clearly presenting this as a limitation or configurable choice. In a marketing audit context, this can silently produce incomplete or misleading results for non-US or global brands, creating integrity risk in downstream reporting and decisions based on the collected data.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.