Back to skill

Security audit

B2B官网SEO全维度审计

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent B2B SEO audit tool, but its generated HTML reports can execute unsafe injected or third-party JavaScript when opened.

Install only if you are comfortable with a Chinese-language SEO audit skill that fetches your site and likely competitor sites and writes report files locally. Treat generated HTML reports as untrusted: avoid opening them in a privileged browser profile or shared origin until the template escapes all website-derived content, validates JSON/chart data, and removes or integrity-pins the remote CDN script.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (2)

T09 · Insecure Skill Coding Practices

Error
Location
assets/html-report-template.html:119
Finding

Unescaped template substitution permits script injection in generated HTML reports

Content
View full analysis
=80: #059669, score>=60: #d97706, score<60: #dc2626) - `{{TECH_GRADE}}` 等 — 评分等级(优秀/良好/需改进/差) - `{{CORE_FINDING}}` — 核心发现(HTML 格式段落) - `{{TECH_ISSUES}}` — 技术问题列表(HTML 格式) - `{{TECH_RADAR_DATA}}` / `{{TECH_BENCHMARK_DATA}}` — 技术雷达图数据(JSON 数组) - `{{CONTENT_RADAR_DATA}}` / `{{CONTENT_BENCHMARK_DATA}}` — 内容雷达图数据(JSON 数组) - `{{CONTENT_MATRIX}}` — 内容覆盖度矩阵(HTML 表格行) - `{{COMPETITOR_HEADERS}}` — 竞品列表表头 - `{{COMPETITOR_ROWS}}` — 竞品对比数据行 - `{{OUR_SITE}}` — 我方简称 - `{{QUICK_WINS}}` / `{{STRATEGIC_KW}}` / `{{FILLER_KW}}` / `{{DEPRIORITIZED_KW}}` — 四象限关键词(HTML li) - `{{TOP20_ROWS}}` — TOP 20 优化表格行 - `{{PHASE1_ITEMS}}` ~ `{{PHASE4_ITEMS}}` — 各阶段任务项 - `{{L1_METRICS}}` / `{{L2_METRICS}}` / `{{L3_METRICS}}` — KPI 指标(HTML 格式) ``` The template then inserts these values directly into HTML contexts: ```html
{{SITE_URL}} — 搜索流量增长策略与执行路线图

💡 核心发现

{{CORE_FINDING}}

{{TECH_ISSUES}}
{{CONTENT_MATRIX}}
买家旅程阶段内容类型现有内容数覆盖率质量评分差距
{{COMPETITOR_HEADERS}}
对比维度{{OUR_SITE}}
Remediation
View remediation
` - `` - JavaScript expression-breaking strings - `javascript:` URLs 10. Document the trust boundary and require generated reports to contain only sanitized, locally controlled markup. ]]>

T08 · Insecure Dependencies

Warning
Location
assets/html-report-template.html:7
Finding

Generated reports execute a third-party CDN script without integrity verification

Content
View full analysis
``` ### Technical Analysis Every generated interactive report downloads and executes JavaScript from a third-party CDN when opened. Although the URL identifies Chart.js version `4.4.0`, the script element does not include a Subresource Integrity hash. The template also lacks a Content Security Policy restricting script execution and outbound connections. As a result, the runtime behavior of an otherwise static report depends on content delivered outside the audited project. A compromise affecting the CDN, its package storage, the upstream artifact, or the delivery path could cause generated reports to execute JavaScript that was not present during the Skill audit. This also reduces reproducibility and offline reliability: the report may behave differently or lose visualization functionality depending on external availability. ### Attack Path 1. The Skill generates an HTML report containing the remote Chart.js script reference. 2. The user opens the report while network access is available. 3. The browser requests the script from `cdn.jsdelivr.net`. 4. If the delivered resource has been maliciously altered through an upstream or CDN supply-chain compromise, the browser has no integrity hash with which to detect the change. 5. The browser executes the altered script in the generated report's context. 6. The malicious script can inspect or modify the report and make network requests permitted by the browser environment. ### Impact Assessment A successful dependency-delivery compromise could provide arbitrary JavaScript execution in every generated report opened while the malicious resource is served. The code could: - Read and modify the report's contents. - ...[truncated 379 chars]
Remediation
View remediation
``` 4. Obtain the integrity value from a trusted build or compute it directly from the reviewed artifact; do not use an unverified hash copied from an unrelated source. 5. Add a restrictive Content Security Policy limiting scripts to the minimum approved source. Where possible, move the template's inline script into a reviewed local file so `unsafe-inline` is unnecessary. 6. Define a controlled dependency-update process that includes source verification, vulnerability review, checksum updates, and report regression testing. 7. Consider generating charts without runtime JavaScript, such as pre-rendered SVG or static images, when report interactivity is not essential. ]]>
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (8)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill advertises many broad trigger phrases such as 'SEO审计', '官网SEO', and 'SEO分析报告', which can cause automatic activation in contexts where the user did not clearly consent to a multi-step website and competitor analysis workflow. In this skill, activation is tied to web fetching, competitor discovery, and report generation, so overbroad triggers increase the chance of unintended network access and data collection.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The activation section lists general intents like improving search traffic or making an SEO execution plan, but does not define boundaries for when the skill should not run or when lighter-weight advice is more appropriate. Because the workflow performs external fetching and automatic competitor identification, unclear activation criteria can lead to unnecessary analysis of third-party sites and actions beyond user expectations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill states that if the user does not provide competitors, it will automatically identify 3-5 competitors via web search, but it does not warn the user that third-party websites will be fetched and analyzed. This matters because users may not expect automated scanning of external domains, and in some environments this can create policy, privacy, or reputational concerns.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The report output section instructs the skill to write a Markdown audit file into the project working directory and generate an HTML artifact, but the user is not warned that files will be created. Unannounced writes can clutter the workspace, overwrite expected outputs, or expose analyzed URLs and findings in shared project directories.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The HTML declares lang="zh-CN", and the visible report content is written entirely in Simplified Chinese, which imposes a specific language/locale on all users. There is no indication in the template that language selection is optional or that the skill is explicitly limited to a China-specific reporting context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

SQP-3 applies to all file types and covers natural-language policy violations such as forcing a specific language without user opt-in. This markdown file presents all substantive guidance in Chinese and does not indicate that the user can choose another language or that the content is intentionally limited to a Chinese-speaking audience.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown skill/reference content forces a specific language for all users through its headings, instructions, and templates. Under the policy rule, a language-specific constraint should either be optional for the user or clearly justified as region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
72% confidence
Finding

The manifest description is written to prescribe the skill in Chinese and does not indicate that users may choose another language for interaction or outputs. For a general-purpose SEO audit skill, forcing a single language without opt-in can violate language/locale policy expectations.

Content

No source excerpt is available for this finding.