ClawHub
Back to skill

Security audit

feasibility-review

Security checks across malware telemetry and agentic risk

Overview

The skill performs a disclosed feasibility-report review workflow that reads user-provided documents, searches external patent/paper/policy sources, and writes a local HTML report.

Install only if you are comfortable using PatSnap/MCP and web-search services with extracted information from the uploaded feasibility report. Avoid using confidential or unpublished project materials unless your organization allows that third-party processing.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
该技能明确要求将完整脚本写入会话目录并生成/删除多个本地文件，但未声明相应权限边界或最小化访问范围。权限与实际能力不透明会导致代理在用户不知情的情况下进行文件读写，若被误触发或被后续提示注入影响，可能造成会话文件被覆盖、持久化敏感内容或意外删除临时数据。

Missing User Warnings

Medium
Confidence
87% confidence
Finding
技能要求读取用户上传文档，并基于提取内容调用专利、论文和网页检索工具，但未明确告知哪些文本片段会被发送到外部服务。对于可研报告和项目指南这类常含未公开技术方案、项目名称、合作单位信息的材料，这种不透明的数据外发可能导致商业敏感信息泄露或超出用户预期的第三方处理。

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.