ClawHub

Baidu Hot Cn

Security checks across malware telemetry and agentic risk

Overview

This skill is a small Baidu hot-topics helper with no credential, persistence, or exfiltration behavior, but users should treat its current output as mock data rather than live trends.

Install only if you are comfortable with a demo-like skill: it currently returns simulated Baidu hot topics, not verified live Baidu data. Do not rely on its rankings, counts, or trend analysis for current events or business decisions unless the publisher updates it to fetch and label real sources clearly.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
79% confidence
Finding
The skill appears to rely on network access to retrieve Baidu hot-search data, but no explicit permissions are declared in the manifest metadata. Undeclared network capability reduces transparency and can bypass user/operator expectations about what external access the skill requires, increasing review and governance risk.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill metadata and function naming claim live Baidu hot-topic, trend, and keyword-popularity retrieval, but the implementation only returns hardcoded mock data. This is a deceptive capability mismatch: downstream users or agents may make decisions based on fabricated data while believing it is current external information.

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The module docstring states it fetches Baidu hot searches, but the code path serves simulated data instead. Misleading documentation increases the likelihood that users, agents, or orchestration systems will trust the output as authoritative, creating integrity risks even without direct code execution abuse.

Vague Triggers

Medium
Confidence
85% confidence
Finding
The trigger list includes very broad terms such as “百度”, “热搜”, and “baidu”, which are likely to appear in ordinary conversation and can cause unintended invocation. Accidental activation can lead to unexpected network requests, context switching, or disclosure of user intent to external services when the user did not mean to use this skill.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal