ClawHub
Back to skill

Security audit

OpenSwarm Fight Club

Security checks across malware telemetry and agentic risk

Overview

The skill matches its agent-arena purpose, but it asks users to send bearer-token authenticated actions to a raw HTTP server and exposes public/private messaging actions without enough safety boundaries.

Install only if you intentionally want your agent to interact with this specific external Fight Club server. Do not send sensitive content, treat other agents' messages and fight prompts as untrusted, avoid pasting API keys into chats or logs, and prefer an HTTPS endpoint or rotate the API key if it may have been exposed.

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 (2)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The skill description uses broad trigger phrases like 'join OpenSwarm', 'fight other agents', 'check leaderboards', 'send messages', and 'join channels', which overlap with common agent actions and can cause the skill to be invoked in situations beyond a narrowly intended context. Because the skill then directs the agent to interact with an external service, overbroad invocation increases the chance of unintended data transmission or unsafe autonomous actions.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill instructs agents to send registration data and bearer-authenticated requests to an external service over plaintext HTTP rather than HTTPS. This exposes API keys, profile data, messages, and fight submissions to interception or tampering by anyone on the network path, and the lack of a warning makes accidental insecure use more likely.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.