ClawHub

阿里云云效mcp对接

Security checks across malware telemetry and agentic risk

Overview

This appears to be a real Yunxiao helper, but it needs review because it can post comments to shared work items and runs an unpinned external npm MCP server with the user's token.

Install only if you trust the external Alibaba Cloud DevOps MCP package and can use a least-privilege Yunxiao token. Set `YUNXIAO_ORG_ID` explicitly, avoid exposing unrelated secrets in the environment, and require explicit user approval before any `create_comment` action.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill explicitly documents a `create_comment` command that performs a write operation against a remote Yunxiao system, but it does not warn that this action modifies shared project data or recommend confirmation before use. In an agent setting, missing guardrails around state-changing operations can lead to unintended comments being posted to production work items, causing workflow disruption, audit noise, or accidental disclosure.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The CLI exposes a write-capable operation (`create_comment`) that posts directly to a remote Yunxiao work item using the configured access token, but it performs no confirmation, dry-run, or safety interlock before modifying project data. In an agent/skill context, this increases the chance of unintended state changes from prompt injection, argument confusion, or mistaken automation, especially because comments can leak sensitive content or spam production work items.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal