ClawHub

Rotifer Arena

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Rotifer evaluation helper, but users should only invoke it for intended Rotifer/Gene/Arena workflows.

Install this only if you intend to evaluate Rotifer Genes or Skills. Before running it, confirm the exact local paths or ClawHub slugs being evaluated, avoid sensitive code unless you understand what the Rotifer CLI submits or imports, and treat the npx Rotifer packages as external code you are choosing to trust.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The example trigger phrases are broad and overlap with common user requests such as "compare" or "which is better," which can cause the skill to activate in situations the user did not intend. In this skill, unintended activation is more concerning because it can lead to importing, compiling, matching opponents, and submitting artifacts to an external evaluation system without sufficiently explicit user consent or scope confirmation.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The workflow states that the skill will compile a Gene, auto-find an opponent, and submit it to the Arena, but it does not warn users that their provided code, files, or other artifacts may be sent to an external service for evaluation. This is dangerous because users may unknowingly disclose proprietary, sensitive, or regulated material, and the auto-submission behavior increases the chance of accidental data exfiltration beyond the local environment.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger list is unusually broad and includes generic comparison terms like "compare," "evaluate," and "which is better," which can cause the skill to activate outside its intended Rotifer-specific context. This creates an over-broad routing surface that may lead an agent to perform unintended CLI-driven evaluation workflows, increasing the chance of unnecessary command execution or confusing task hijacking.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill instructs the agent to write a report file when the user replies "save," but it does not clearly warn that this will modify the local filesystem or require confirmation of the exact target path. In an agentic environment, implicit file writes can surprise users, overwrite existing files, or be abused through crafted gene names or paths if filename components are not sanitized.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal