ClawHub
Back to skill

Security audit

协作平台卡片基础版

Security checks across malware telemetry and agentic risk

Overview

This skill is a simple Feishu card-sending guide with expected external messaging behavior, but it should be used carefully because messages and optional callback data leave the local environment.

Before installing or using it, verify the Feishu target ID, review message text for secrets or personal data, and avoid using callback_url unless you trust the endpoint and understand what will be sent. Also inspect the separately installed feishu-common and send.js/send_safe.js code, because this artifact relies on them but does not include them.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
91% confidence
Finding
The skill accepts a `callback_url` for asynchronous notification even though its stated purpose is basic card sending. This creates an unnecessary outbound data path that could exfiltrate message content, metadata, or execution status to arbitrary external endpoints, especially in agent environments where users may not realize callbacks transmit data off-platform.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill documents outbound messaging and callback behavior without a prominent warning that user-provided content and identifiers are transmitted to external services. In practice, this can lead to privacy leaks, accidental disclosure of internal data, or misuse in automated workflows where operators assume the action is local.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The examples encourage sending content to user and group targets but omit a user-facing warning that this causes immediate external delivery. In an agent context, example commands are often copied directly, so the lack of disclosure increases the chance of accidental transmission of logs, alerts, or sensitive operational details to unintended recipients.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.