ClawHub
Back to skill

Security audit

水滴筹众筹助手

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed ShuiDiChou donation and fundraising helper, but users should be careful before sharing a phone number or donating.

Install only if you want an assistant that can contact ShuiDiChou APIs, show donation links, and submit a phone number for fundraising follow-up after you provide it. Verify the official domain and affiliation before donating or sharing contact details, and treat broad medical or charity prompts as a cue to confirm intent before proceeding.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill invokes shell scripts (`donate.sh` and `register.sh`) but does not explicitly declare corresponding permissions or execution capabilities. This creates a transparency and governance gap: reviewers and users cannot accurately assess what code will run or what external actions may occur, especially since one path submits a user's phone number to an external API.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The donation trigger list contains broad phrases such as '爱心' and '做好事' that can match ordinary conversation unrelated to crowdfunding. Overbroad activation can cause the skill to engage unexpectedly, steering users toward donation links they did not request and increasing the risk of unwanted persuasion or misfire in sensitive contexts.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The fundraising trigger list includes ambiguous hardship phrases like '家里有人生病' and '看不起病', which may activate the skill during emotionally sensitive medical conversations without a clear request to use this service. In this skill's context, unintended activation is more dangerous because it can prompt collection of a phone number and transmission of personal data to an external platform.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The script sends a user's phone number to a third-party API endpoint, which is personally identifiable information, without any built-in consent prompt, disclosure, or verification that the user explicitly agreed at execution time. In this skill's context, the function is specifically to submit fundraising leads, so silent transmission of contact data is privacy-sensitive and can result in unauthorized sharing of medical-adjacent personal information.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal