ClawHub
Back to skill

Security audit

Fuzzy Multi Agent Team

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed OpenClaw agent-and-Discord setup helper, but it can change agent configuration, create persistent workspaces, and optionally schedule ongoing agent activity.

Before installing, review the gateway patch commands, use a dedicated workspace instead of a broad home or project directory, grant Discord permissions only as needed, and enable cron or long-running agents only when you want persistent automation.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The skill description is written with very broad trigger language such as "multi-agent," "parallel agents," and "coordinated agents," which can cause the skill to activate for common planning or coordination requests rather than only explicit sub-agent orchestration tasks. In a system that can spawn autonomous workers, over-broad matching increases the chance of unintended agent creation, unnecessary delegation, and execution of higher-risk behaviors without clear user intent.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly instructs persistent agents to update shared workspace files, but it does not require user awareness, approval, path restrictions, or safeguards around what files may be modified. In a multi-agent setting this is more dangerous because several agents may concurrently write state, overwrite unrelated files, leak sensitive context into shared artifacts, or create durable side effects that the user did not intend.

Static analysis

No suspicious patterns detected.