ClawHub
Back to skill

Security audit

Alibabacloud Find Skills

Security checks across malware telemetry and agentic risk

Overview

This skill is mainly a marketplace search helper, but its install workflow can persistently add other skills globally while skipping confirmation.

Review carefully before installing. Use the discovery/search parts freely for Alibaba Cloud skill lookup, but do not run the global `npx ... -g -y` examples unless you intentionally want to add persistent skills to your agent. Prefer interactive or project-local installation, review the selected skill source first, and avoid putting secrets, credentials, personal data, or internal project details in search queries.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (7)

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The skill claims to be read-only and to create no resources, yet its workflow includes `npx` installation commands that modify the local environment and may install code from external packages. This mismatch can mislead users or calling agents into treating the skill as non-invasive, reducing scrutiny before executing environment-changing actions.

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest positions the skill as a search/discovery tool, but the documented workflow extends into installation and execution of other skills. That scope expansion increases the chance an orchestrator invokes the skill in contexts where only read-only discovery was expected, leading to unanticipated code installation or follow-on actions.

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger phrases include broad terms like '阿里云', '帮我找一个 skill', and '使用阿里云相关的skill', which can match ordinary conversation and cause unintended invocation. In this skill, accidental invocation is more risky because the workflow can progress from search into installation guidance or execution of downstream skills.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill provides `npx skills add` and `npx clawhub install` commands, including global and non-interactive options, without strong warnings about side effects such as modifying the home directory, installing remote packages, or trusting package registries. Users or agents may execute these commands without understanding that they introduce persistent code and configuration changes.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The examples instruct direct requests to an external Alibaba Cloud API using user-derived search phrases, but they do not warn that the user's query text will be transmitted off-platform. In a skill whose purpose is discovering skills from a remote marketplace, this increases privacy and data-handling risk because users or downstream agents may include sensitive operational details, internal project names, or other confidential text in search terms without informed consent.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The document instructs users to run a global `npx skills add ... -g -y` installation command that can modify the local environment, but it does not warn about the system-wide effects, trust boundary, or need to review the package and skill source first. In a skill-discovery context, this increases the chance that a user treats discovery and installation as equally safe and executes a command that changes their workstation or agent runtime without adequate caution.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
# Recommended: use non-interactive mode to avoid blocking.
#   --agent <client>   Agent client to install for (see references/npx-skills-agents.md)
#   -g                 Install globally (home dir); omit for project-local install
#   -y                 Skip confirmation (requires --agent and -g/-local to be set)
npx skills add aliyun/alibabacloud-aiops-skills \
  --skill <skill-name> \
  --full-depth \
Confidence
98% confidence
Finding
Skip confirmation

VirusTotal

60/60 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.