ClawHub
Back to skill

Security audit

Article Fetcher（文章抓取+Notion/Obsidian知识库存档）

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed article-archiving skill with real privacy considerations, but I found no hidden, unrelated, destructive, or deceptive behavior.

Install only if you are comfortable with article images being uploaded to Aliyun OSS, optional Notion archiving, and optional LLM tag extraction sending article text to your configured LLM provider. Use least-privilege OSS and Notion credentials, keep WeChat/Zhihu cookie files narrowly scoped, and prefer a low-risk account for cookie-based scraping.

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 (4)

Intent-Code Divergence

Medium
Confidence
88% confidence
Finding
The README asserts that cookies and other sensitive information 'will not be uploaded or leaked', yet the documented workflow explicitly uses authenticated cookies and a Playwright browser session to access remote third-party sites. Even if the skill does not intentionally exfiltrate cookies, this claim is overly broad and misleading because authenticated state is necessarily transmitted to target platforms and may be exposed through browser automation, request headers, logs, crash artifacts, or misconfiguration.

Missing User Warnings

Medium
Confidence
85% confidence
Finding
The Playwright fallback automatically injects all stored cookies into a browser context for the target URL, which can silently authenticate requests and expose session-bound data without an explicit trust boundary or user confirmation at the point of use. In a scraping skill that fetches arbitrary article URLs, this increases the risk of unintended credential use, privacy leakage, or sending sensitive cookies to broader domains than expected if cookie scope is not tightly validated.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The Playwright fallback copies all stored cookies into a browser context and sends them to Zhihu automatically, which can reuse authenticated session state without any explicit user-facing consent or scope restriction in this file. In a scraping tool that archives content from third-party platforms, this increases privacy and account-risk exposure because sensitive session cookies may be used in an automated browser flow and could trigger account actions, tracking, or unintended authenticated access.

External Transmission

Medium
Category
Data Exfiltration
Content
for attempt in range(3):
        try:
            timeout = [60, 90, 120][attempt]
            response = requests.post(endpoint, headers=headers, json=body, timeout=timeout)

            if response.status_code == 200:
                return response.json()['choices'][0]['message']['content']
Confidence
83% confidence
Finding
requests.post(endpoint, headers=headers, json=

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.