ClawHub
Back to skill

Security audit

自集成工具

Security checks for vulnerabilities and agentic risk

Overview

This skill is an external-app integration guide that can perform live actions in connected services, but it does not clearly require user confirmation or scoping before writes, deletes, or automated cross-app actions.

Review this before installing if you plan to connect real Slack, HubSpot, Notion, CRM, or business accounts. Use a limited gateway token, test with non-production accounts first, and require explicit approval before any create, update, delete, send, or webhook-triggered action.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The skill claims the free edition does not support workflow automation, yet the usage example describes Slack-triggered synchronization into HubSpot, which effectively instructs agents to build or simulate an automated cross-application workflow. This mismatch can cause agents or users to over-trust unsupported behavior and attempt unintended data-moving automations across external systems.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The skill's activation guidance is broad ('use when API integration, webhook configuration, or system connection is needed'), which can cause an agent to invoke a high-privilege integration skill in many loosely related situations. Because the skill can connect to external services and execute actions, vague routing criteria increase the risk of unnecessary external calls or unintended side effects.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The description advertises the ability to execute actions against external applications but does not prominently warn that those actions may create, modify, send, or delete data in third-party systems. In agent settings, this omission is risky because users may interpret the skill as informational rather than operational, leading to unintended real-world changes.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The documentation explicitly states support for create/query/modify/delete-style operations and shows action execution flows, but it lacks a mandatory confirmation step for destructive or data-modifying operations. In a skill intended for agents with exec/network capability, this materially increases the chance of accidental writes or deletions in connected third-party applications.

Static analysis

No suspicious patterns detected.