ClawHub

ClawHub 中文 Skill 查询

Security checks across malware telemetry and agentic risk

Overview

This is a public ClawHub search helper with an optional, disclosed daily notification feature that users should enable deliberately.

For normal one-time searches, this skill only queries public ClawHub data. Before enabling daily reports, confirm the schedule, recipient/channel, and how to list or remove the cron job so recurring messages do not continue unexpectedly.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Description-Behavior Mismatch

Medium
Confidence
92% confidence
Finding
The skill’s stated purpose is an on-demand query for newly uploaded Chinese skills, but it also instructs the agent to create persistent cron jobs that proactively push results to a user channel. This expands behavior from user-requested retrieval into ongoing autonomous messaging, which can enable unwanted persistence, spam, or monitoring beyond the original request if delivery targets are misconfigured or user consent is unclear.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
The skill grants broad cron-management capabilities including add, update, remove, list, and run, plus delivery targeting to user channels, even though its main purpose is simply filtering Chinese-language skills from a webpage. That excess capability increases the risk of abuse, accidental reconfiguration, unauthorized message delivery, or persistence mechanisms that are not necessary for the core task.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal