ClawHub
Back to skill

Security audit

RSS AI 摘要 LITE

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed RSS summarizer that sends feed summaries to an LLM and Feishu, with some documentation gaps users should notice before use.

Use this only for feeds you are comfortable sending to the configured LLM service and Feishu group. Keep API keys and webhook URLs in secure environment variables, review any referenced Python project and requirements before running, and treat the callback_url field as unsupported unless the publisher documents exactly how it is used.

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 (3)

Intent-Code Divergence

Medium
Confidence
86% confidence
Finding
The skill advertises Feishu-only push for the LITE version, yet its documented input accepts an arbitrary callback_url for asynchronous notifications. That creates an undocumented outbound communication path that could be used to exfiltrate processing results or metadata to attacker-controlled endpoints, especially in agent environments where users may trust the stated channel restrictions.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill clearly describes sending feed content to an LLM backend and summaries to Feishu, but it does not prominently disclose that external services will receive potentially sensitive feed data, article content, URLs, and derived summaries. In an agent context, weak disclosure increases the risk of unintentional data leakage because users may treat the workflow as local automation rather than third-party transmission.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The documentation tells users to provide API keys and webhook URLs through environment variables but does not warn that these are sensitive secrets. In practice, users may place them in shared configs, shell history, logs, screenshots, or committed files, leading to credential leakage and abuse of the LLM account or webhook endpoint.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.