ClawHub

魔方 Data 轻量版问题解答

Security checks across malware telemetry and agentic risk

Overview

This skill is a transparent staff moderation helper with powerful actions, but its instructions are scoped, disclosed, and require authenticated, audited tooling.

Install only if you are a ClawHub staff operator or maintainer expected to perform moderation. Treat ban, unban, role, and unhide commands as high-impact actions: confirm the target and reason, rely on the normal authenticated CLI/API path, and verify audit-backed results after each write.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Context-Inappropriate Capability

High
Confidence
98% confidence
Finding
The skill embeds a default username and password in plain text, which exposes credentials to any user or downstream system that can read the skill. In a FAQ/support skill, this is especially risky because it normalizes credential disclosure and can enable unauthorized access if the credentials are valid in deployed environments.

Ssd 3

High
Confidence
99% confidence
Finding
The file explicitly instructs users to use default login credentials, disclosing a password in plain language. If these credentials are active or reused, attackers or unauthorized users could log into the system; even if they are starter credentials, publishing them in a broadly accessible support artifact materially increases the risk of compromise.

VirusTotal

51/51 vendors flagged this skill as clean.

View on VirusTotal