Back to skill

Security audit

Brand Visibility Monitor

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a brand visibility monitor, but it needs Review because it sends business search data to external services and ships weak API/licensing controls.

Install only if you are comfortable sending brand names, search prompts/results, license checks, and optional reports to external AI platforms, yk-global, Tavily, and Feishu. Do not deploy the bundled API service as-is; replace the placeholder key, add real auth and rate limits, and fix the local upgrade and temporary-report handling first.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
scripts/geo_quota.py:274
Finding

Unauthenticated local upgrade commands bypass paid-tier quota restrictions

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
api/geo_api.py:21
Finding

Repository-known placeholder key authenticates to the API service

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/geo_report.py:139
Finding

Predictable shared temporary report path permits symlink overwrite

Content
View full analysis
Remediation
View remediation
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
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (48)

Tainted flow: 'TAVILY_API_KEY' from os.environ.get (line 15, credential/environment) → requests.post (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · api/geo_api.py (reported line 34)May include surrounding context.

python
def search_brand_tavily(brand_name, max_results=5):
    """使用Tavily搜索品牌信息"""
    try:
        response = requests.post(
            TAVILY_API_URL,
            json={
                "api_key": TAVILY_API_KEY,

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

This second mismatch finding also points to materially misleading capability claims, especially around AI-based causal analysis, webhook pushing, and competitor monitoring. In an agent ecosystem, overstated automation and hidden reliance on external systems can lead to unintended disclosure, improper operator trust, and unsafe business decisions based on incomplete or fabricated functionality.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding

This second mismatch finding also points to materially misleading capability claims, especially around AI-based causal analysis, webhook pushing, and competitor monitoring. In an agent ecosystem, overstated automation and hidden reliance on external systems can lead to unintended disclosure, improper operator trust, and unsafe business decisions based on incomplete or fabricated functionality.

Content

No source excerpt is available for this finding.

Lp1

High
Category
MCP Least Privilege
Confidence
89% confidence
Finding

This service performs outbound requests to Tavily using user-supplied brand queries, but that network capability is not reflected in the declared permissions. Undeclared external transmission is dangerous because users and reviewers cannot accurately assess data exposure or constrain the skill's runtime behavior.

Content

No source excerpt is available for this finding.

Lp1

High
Category
MCP Least Privilege
Confidence
89% confidence
Finding

This service performs outbound requests to Tavily using user-supplied brand queries, but that network capability is not reflected in the declared permissions. Undeclared external transmission is dangerous because users and reviewers cannot accurately assess data exposure or constrain the skill's runtime behavior.

Content

No source excerpt is available for this finding.

Lp1

High
Category
MCP Least Privilege
Confidence
87% confidence
Finding

The skill has shell-capable behavior through subprocess execution, but this capability is not declared in permissions. Undeclared execution capability is dangerous because it weakens review transparency and can allow networked command execution paths that operators and users did not explicitly approve.

Content

No source excerpt is available for this finding.

Lp1

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'file_read' capability that is not listed in its permissions. This may indicate deceptive intent or missing permission declarations.

Content

No source excerpt is available for this finding.

Lp1

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'file_write' capability that is not listed in its permissions. This may indicate deceptive intent or missing permission declarations.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The documentation states that paid tiers are unlocked through API verification, but the implementation also supports local tier escalation by mutating the quota file. This discrepancy can mislead reviewers and operators into trusting a licensing control that is easily bypassed, weakening control integrity and making abuse more likely.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The module exposes local functions that directly mutate the quota state to 'pro' or 'enterprise' without any API-key verification or server-side authorization. Any caller able to invoke these functions, or to edit the quota file equivalently, can bypass licensing restrictions and unlock unrestricted execution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The README explicitly advertises a server-side relay and Feishu webhook integration but does not explain what brand queries, prompts, results, or identifiers may be transmitted, stored, or retained. In a skill that monitors brand visibility across third-party AI platforms, this omission can cause users to unknowingly send potentially sensitive business intelligence or internal brand-monitoring data to external services.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The triggers are broad and generic for a skill that can invoke local scripting and potentially perform external requests, increasing the chance of accidental activation in unrelated brand, AI, or monitoring conversations. Unintended invocation can cause unnecessary data collection, quota consumption, external transmission, or execution of automation without sufficiently specific user intent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill documents outbound API use, server-side relay, and webhook pushing without a clear, prominent user warning about external data transmission. This is dangerous because brand names, queries, API keys, and report contents may be sent to third parties or custom webhooks without users fully understanding the data path and retention implications.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

The documented external endpoint and server relay indicate intentional off-host data transmission to a vendor-controlled API. While external communication is not inherently malicious, it is security-relevant here because the skill handles user-supplied brand intelligence inputs and license/API credentials, and the relay architecture reduces local visibility into what is transmitted or retained.

Content

Scanner excerpt · SKILL.md (reported line 41)May include surrounding context.

md
- ✅ 我们服务器中转,无需本地环境
- ✅ 稳定高效,不封IP

> 购买专业版/企业版后,请在 [https://yk-global.com](https://yk-global.com) 获取API密钥,使用 `--api-key YOUR_KEY` 参数运行即可解锁全部功能。验证接口:`POST https://api.yk-global.com/v1/verify`,验证失败自动降级到免费版,不阻断使用。

## 支持的AI平台

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This code includes natural-language content exclusively in Chinese in the module description and runtime error messages, with no indication that the skill is region-specific or that users can opt into the locale. That creates a language policy concern because the skill effectively mandates a specific language for operators and API consumers.

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 · api/geo_api.py (reported line 18)May include surrounding context.

python
TAVILY_API_KEY = os.environ.get('TAVILY_API_KEY', '')
if not TAVILY_API_KEY:
    raise ValueError("请设置 TAVILY_API_KEY 环境变量")
TAVILY_API_URL = "https://api.tavily.com/search"

# API认证(简易版:检查X-API-Key头)
API_KEYS = {

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest describes automatic visibility checks on several specific AI platforms and generation of analysis about why a brand was not recommended. In this file, the implemented behavior is a single POST request to Tavily with a generic query, which is materially narrower than the advertised platform-specific monitoring capability.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
82% confidence
Finding

The endpoint sends user-provided brand queries and the service's API credential to an external third party (Tavily). In the context of a brand-monitoring skill, external transmission is expected, but it is still security-relevant because user inputs may contain sensitive business terms and are disclosed to a provider outside the local trust boundary.

Content

Scanner excerpt · api/geo_api.py (reported line 34)May include surrounding context.

python
def search_brand_tavily(brand_name, max_results=5):
    """使用Tavily搜索品牌信息"""
    try:
        response = requests.post(
            TAVILY_API_URL,
            json={
                "api_key": TAVILY_API_KEY,

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest promises higher-level reasoning and notification features, but the code computes a simplified score solely from result count and exposes no analysis logic or Feishu push behavior. This is a meaningful mismatch between claimed product behavior and implemented functionality in this file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This Python file contains user-facing natural-language strings entirely in Chinese, including the module docstring, function docstring, and console output. Under the policy rule for natural-language violations, forcing a specific language without user opt-in or documented regional justification is a locale-policy issue.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
80% 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 · api/test_api.py (reported line 11)May include surrounding context.

python
def test_search(brand="91tokenhub"):
    """测试搜索"""
    resp = requests.post(
        API_URL,
        headers={"X-API-Key": TEST_KEY},
        json={"brand": brand, "max_results": 5}

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The config sets the report language to "zh-CN" unconditionally. This is a natural-language locale constraint, and the file does not indicate any user opt-in, alternative locale selection, or region-specific justification for forcing Chinese output.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The module description, prompt, fallback analysis, and CLI usage strings are all written exclusively in Chinese, which effectively fixes the skill's interaction language. The file does not offer a language choice or explain that the skill is intentionally region-specific, so this is a locale policy concern.

Content

No source excerpt is available for this finding.