ClawHub
Back to skill

Security audit

xinjianxue-meihua

Security checks for vulnerabilities and agentic risk

Overview

The skill provides a clearly priced divination service with buyer-controlled payment, while sending your numbers and optional question to an external provider.

Expect a 6 CNY charge per delivered reading and separate Alipay authorization. Verify the merchant and amount before paying. Your question can reach the provider before payment, so omit names and confidential details; privacy and retention terms are not supplied.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

The workflow sends user-provided inputs to a remote divination service without a privacy notice.

Medium
Category
Output Handling
Confidence
85% confidence
Finding

The workflow collects user inputs and invokes an external service without explaining what information leaves the conversation or who receives it. Although the two numbers alone are generally low sensitivity, the optional question can reveal private relationship, employment, or financial circumstances if transmitted; the unavailable API reference prevents confirming whether that question is actually included. The instruction not to request location reduces unnecessary collection, but does not address disclosure of other personal information.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/api.md (reported line 3)May include surrounding context.

md
# 接口契约

- 地址：`https://api.xinjianxue.com/api/xinjianxue/aipay/meihua`
- 方法：POST，`Content-Type: application/json`
- 单价：6 元/次
- 鉴权：只认 `Payment-Proof`，不需要账号、密码或登录态

The request workflow sends the user's question to an external API without a privacy disclosure.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The request sends the user's question and chosen numbers to an external service without instructing the agent to disclose that transfer or obtain informed permission first. Divination questions can contain sensitive personal, relationship, financial, or business information, and transmission occurs before the checkout disclosure.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/payment.md (reported line 8)May include surrounding context.

1. 发起请求

bash
curl -sS -D - -X POST "https://api.xinjianxue.com/api/xinjianxue/aipay/meihua" \
  -H "Content-Type: application/json" \
  -d '{"n1":8,"n2":23,"question":"这次合作能不能成"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/payment.md (reported line 8)May include surrounding context.

1. 发起请求

bash
curl -sS -D - -X POST "https://api.xinjianxue.com/api/xinjianxue/aipay/meihua" \
  -H "Content-Type: application/json" \
  -d '{"n1":8,"n2":23,"question":"这次合作能不能成"}'

The skill introduces a mandatory paid checkout flow that is not disclosed in its divination-only manifest.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The manifest presents a divination service without disclosing that delivery requires a paid checkout and payment integration. This creates a transparency and informed-consent gap, although the instructions require showing payment details and buyer authorization, so they do not establish unauthorized charging.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.