ClawHub
Back to skill

Security audit

Cms Search

Security checks across malware telemetry and agentic risk

Overview

This is a coherent web-search skill that sends search queries to a remote CMS search service, with no evidence of persistence, local data harvesting, destructive behavior, or hidden actions.

Install only if you are comfortable sending search terms to the CMS search endpoint using the CMS_USER_KEY supplied by your environment. Avoid putting secrets, private personal data, confidential business details, or regulated medical/financial information into queries unless that external transmission is acceptable.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding
The skill invokes a local Python script that requires environment-provided credentials and performs live network access, yet the skill metadata declares no permissions. This creates a transparency and policy-enforcement gap: a host system or reviewer may underestimate what the skill can access, allowing unintended use of secrets or outbound requests.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The activation description is extremely broad and overlaps with generic search, news, policy, market, medical, and financial queries, making the skill likely to trigger for a large share of user requests. Over-broad activation can route sensitive or high-stakes queries into a networked tool unnecessarily, increasing data exposure and reducing user control over when external retrieval is used.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The skill sends user-provided search terms and a credential header to an external web service, but the code provides no disclosure, consent prompt, or privacy warning to the operator. In an agent-skill context, this is risky because users may enter sensitive internal, medical, financial, or competitive information, which would then be transmitted off-box to a remote service without clear notice.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.