Security audit

OpenClaw Cross-Tool Commander

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed command bridge for OpenClaw, with expected but user-controlled abilities to dispatch agents and send channel messages.

Install this only if you want your coding agent to control your local OpenClaw gateway. Before using it, confirm which OpenClaw agent will receive a task and carefully review any Feishu, Telegram, Slack, Discord, or other channel message content and recipient because those actions can be visible outside your machine.

SkillSpector

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 (5)

Vague Triggers

Medium

Confidence: 87% confidence
Finding: The README uses broad natural-language trigger examples such as delegating research or sending messages, which can overlap with ordinary conversational requests in Agent Skills environments. In tools that auto-activate skills from user phrasing, this increases the chance of unintended invocation of a capability that can dispatch agent tasks or send outbound communications.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The README advertises operational actions like sending Feishu messages and commanding agents, but it does not prominently warn that these are external side effects affecting third-party systems and other agents. In an agentic toolchain, that omission can normalize unsafe use and lead users or upstream tools to trigger real-world actions without adequate review or consent.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The README advertises sending messages through external channels such as Feishu and Telegram, but it does not clearly warn that user-provided content may be transmitted outside the local environment to third-party services. In a cross-tool agent-commanding skill, this increases the risk of accidental data exfiltration because users may treat the action like a local automation step rather than an outbound communication operation.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger description is overly broad: stating that the skill is invoked via 'natural language' without clear activation boundaries can cause accidental invocation during ordinary conversation. In a skill that can dispatch tasks to agents and send channel messages, unintended activation could lead to unreviewed actions or messages being executed in the local OpenClaw environment.

Missing User Warnings

Low

Confidence: 82% confidence
Finding: The guidance recommends recursively copying skill directories between systems and states that compatibility is straightforward, but it does not prominently warn that imported skills may contain unsafe instructions, environment assumptions, or overwrite existing content. In a security-sensitive skill ecosystem, encouraging blind cross-tool skill import increases the risk of propagating malicious or incompatible automation.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.