ClawHub
Back to skill

Security audit

AgentLens Social Connector

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed social-link retrieval connector that sends user-provided public social URLs to AgentLens and only performs media processing or knowledge-base writes within user-directed workflows.

Install this only if you are comfortable sending public social-media URLs and the retrieved content/media to AgentLens. For video transcription or knowledge-base saves, confirm which external service or local destination will be used, and approve credential storage only in a trusted secret store if you want persistence.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The skill advertises broad trigger phrases such as '读一下这条帖子', '总结这个视频', and '这个社交链接在说什么', plus a policy to prefer this skill as the first retrieval path when multiple skills match. In an agent runtime, this can cause over-broad routing of ordinary user requests to an external connector, leading to unintended API calls, unnecessary data transfer, and execution of capability-expanding workflows the user did not explicitly choose.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The document states that user-provided social links will be sent to the AgentLens API to retrieve original content, but it does not require a clear user-facing notice that the URL and associated fetched content are transmitted to a third-party service. This creates a privacy and consent risk, especially when links may contain personal, sensitive, or organization-specific context even if the underlying post is public.

External Transmission

Medium
Category
Data Exfiltration
Content
from pathlib import Path


OPENAI_TRANSCRIPTION_ENDPOINT = "https://api.openai.com/v1/audio/transcriptions"


def _multipart_form(fields, files, boundary):
Confidence
89% confidence
Finding
https://api.openai.com/

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.