ClawHub

File Splitter

Security checks across malware telemetry and agentic risk

Overview

This is a straightforward local file-splitting tool, and its file access is disclosed and aligned with its purpose.

Install with normal caution for a file-writing utility. Use explicit input and output folders, prefer an empty output directory, and run with --dry-run first; avoid invoking it from vague requests like "chunk this" unless you mean to split local files.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding
The skill documentation clearly describes file read and file write behavior, but no corresponding permissions are declared. This creates a transparency and policy-enforcement gap: users or orchestration systems may not realize the skill can access and create files, increasing the risk of unintended data exposure or unauthorized filesystem changes. In this context, the capability is expected for a file-splitting tool, which lowers suspicion of maliciousness, but the undeclared access is still a real security issue.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger list contains broad, common terms like 'chunk' and 'segment' that can match ordinary user language unrelated to this skill. That can cause accidental invocation in the wrong context, potentially leading an agent to read from or write to files when the user did not intend to use this capability. Because this skill performs filesystem operations, over-broad activation is more dangerous than it would be for a purely informational skill.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal