ClawHub
Back to skill

Security audit

AI-HIVE 多模态创意工具箱

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI-HIVE image/video generation toolkit, with expected API, local media, and file-output behavior for that purpose.

Before installing, expect this skill to use an AI-HIVE API key, upload only media files you explicitly pass to its commands, potentially incur AI-HIVE generation charges, and save generated outputs locally. Use the documented confirmation step before paid or batch jobs and avoid storing real API keys in shared environments.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill includes executable code paths and operational guidance that use environment variables, local files, shell commands, and network access, but it does not declare those permissions up front. That creates a capability/expectation mismatch: a caller or reviewer may treat the skill as documentation-only while it can actually trigger external API calls, local media processing, and file handling, increasing the chance of unintended data exposure, unreviewed execution, or surprise billing.

Vague Triggers

Medium
Confidence
96% confidence
Finding
The skill enables implicit invocation with no listed trigger phrases, scope boundaries, or confirmation step. That creates a prompt-routing risk where unrelated user requests may automatically activate this skill, causing unintended use of external multimodal generation workflows or code-oriented outputs without clear user intent.

Natural-Language Policy Violations

Medium
Confidence
89% confidence
Finding
The default prompt hard-codes Chinese output and a production-ready workflow format without checking the user's language preference or consent. This can mis-handle user intent, reduce transparency, and increase the chance the agent performs more action-oriented orchestration than the user expected, though the security impact is limited compared with direct code execution flaws.

Static analysis

No suspicious patterns detected.