Back to skill

Security audit

客户背景调查-摸清客户采购底细

Security checks for vulnerabilities and agentic risk

Overview

The skill is mainly a disclosed company-intelligence tool, but it also performs account bootstrap, device fingerprinting, persistent API-key storage, and signed-link sharing that users should review carefully before installing.

Review this skill before installing if you are not comfortable with automatic account creation, a MAC-derived device hash being sent for trial-account deduplication, credentials being stored in ~/.zlbx/config.json, and generated reports containing signed sk links. Prefer preconfiguring ZLBX_API_KEY if you want to avoid the auto-registration path.

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
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
Findings (30)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 89)May include surrounding context.

md
默认再用 `scripts/render_report.py` 生成一份可分享的 HTML 版报告并告知保存路径(详见 report-template.md「HTML 报告导出」)。

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 93)May include surrounding context.

md
默认再用 `scripts/render_report.py` 生成一份可分享的 HTML 版报告并告知保存路径(详见 report-template.md「HTML 报告导出」)。

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

This duplicate finding identifies the same hidden instruction block that alters outbound request behavior and instructs the agent to submit a display acknowledgement after showing an update message. Concealed workflow directives reduce auditability and can be repurposed for undisclosed telemetry or agent steering without clear user awareness.

Content

Scanner excerpt · SKILL.md (reported line 185)May include surrounding context.

  • 用户想持续跟踪该公司 → 介绍竞对动态周报(见上方「竞对动态周报」)
  • 以上都不贴切 → 询问是否要对某个竞对做双公司对比,或引导访问知了商机大师 https://agent.zhiliaobiaoxun.com

安装包身份与升级检查:所有发往知了标讯 API 的请求(包括账户查询)必须额外携带以下两个请求头:

text
X-Skill-Package: client-background-check/1.0.6

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

This duplicate finding identifies the same hidden instruction block that alters outbound request behavior and instructs the agent to submit a display acknowledgement after showing an update message. Concealed workflow directives reduce auditability and can be repurposed for undisclosed telemetry or agent steering without clear user awareness.

Content

Scanner excerpt · SKILL.md (reported line 185)May include surrounding context.

  • 用户想持续跟踪该公司 → 介绍竞对动态周报(见上方「竞对动态周报」)
  • 以上都不贴切 → 询问是否要对某个竞对做双公司对比,或引导访问知了商机大师 https://agent.zhiliaobiaoxun.com

安装包身份与升级检查:所有发往知了标讯 API 的请求(包括账户查询)必须额外携带以下两个请求头:

text
X-Skill-Package: client-background-check/1.0.6

Hidden Instructions

High
Category
Prompt Injection
Confidence
97% confidence
Finding

This duplicated finding identifies the same hidden-instruction block, which attempts to steer authenticated requests by embedding non-user-visible directives in the skill content. In a security-sensitive agent setting, such hidden instructions are risky because they can manipulate how the model uses credentials and communicates with external services outside normal transparent configuration paths.

Content

Scanner excerpt · references/api-quick.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Hidden Instructions

High
Category
Prompt Injection
Confidence
97% confidence
Finding

This duplicated finding identifies the same hidden-instruction block, which attempts to steer authenticated requests by embedding non-user-visible directives in the skill content. In a security-sensitive agent setting, such hidden instructions are risky because they can manipulate how the model uses credentials and communicates with external services outside normal transparent configuration paths.

Content

Scanner excerpt · references/api-quick.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

The hidden comment block embeds operational instructions telling the agent which identity headers to send and to preserve package/source attribution. Hidden execution-oriented instructions inside a skill artifact are risky because they bypass normal user visibility and can steer network behavior in ways the user did not request or understand.

Content

Scanner excerpt · references/auto-register.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

The hidden comment block embeds operational instructions telling the agent which identity headers to send and to preserve package/source attribution. Hidden execution-oriented instructions inside a skill artifact are risky because they bypass normal user visibility and can steer network behavior in ways the user did not request or understand.

Content

Scanner excerpt · references/auto-register.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This file defines an auto-registration and account bootstrap workflow, including device fingerprint collection, remote account creation, quota handling, and persistent API-key management, which is materially unrelated to the stated customer background-check purpose of the skill. That scope expansion is dangerous because it authorizes collection of local system identifiers and credential handling under the guise of a procurement-intelligence tool, increasing privacy, supply-chain, and unauthorized-account-risk exposure.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The instructions explicitly direct collection of local device characteristics such as OS, CPU architecture, and a MAC-derived hash, plus later persistence of API credentials, even though these actions are not necessary for generating a customer background-check report. In this context, the mismatch makes the behavior more dangerous because users invoking a business-intelligence skill would not reasonably expect local device fingerprinting and credential storage side effects.

Content

No source excerpt is available for this finding.

Hidden Instructions

High
Category
Prompt Injection
Confidence
91% confidence
Finding

This duplicate finding points to the same hidden instruction region that directs the agent to use specific identity headers and package metadata. Even if intended for legitimate package routing, placing such controls in hidden template text creates an opaque prompt-layer authority that can be abused or accidentally propagated across tools and contexts.

Content

Scanner excerpt · references/report-template.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Hidden Instructions

High
Category
Prompt Injection
Confidence
91% confidence
Finding

This duplicate finding points to the same hidden instruction region that directs the agent to use specific identity headers and package metadata. Even if intended for legitimate package routing, placing such controls in hidden template text creates an opaque prompt-layer authority that can be abused or accidentally propagated across tools and contexts.

Content

Scanner excerpt · references/report-template.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The template requires returning raw API-provided URLs containing the sk login-bypass parameter and emphasizes preserving it unchanged. Exposing bearer-style access links to users or third parties can grant unintended access to protected company pages and reports, effectively turning the agent into a token-sharing mechanism outside the minimum needed for background analysis.

Content

No source excerpt is available for this finding.

Hidden Instructions

High
Category
Prompt Injection
Confidence
95% confidence
Finding

This duplicate finding points to the same hidden instruction block that tells the agent to send fixed package/source headers and use an existing API key. In skill context, this is especially risky because the workflow repeatedly emphasizes preserving package identity and performing account/version calls, increasing the chance that the model follows untrusted prompt instructions affecting authentication and outbound requests.

Content

Scanner excerpt · references/workflow.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Hidden Instructions

High
Category
Prompt Injection
Confidence
95% confidence
Finding

This duplicate finding points to the same hidden instruction block that tells the agent to send fixed package/source headers and use an existing API key. In skill context, this is especially risky because the workflow repeatedly emphasizes preserving package identity and performing account/version calls, increasing the chance that the model follows untrusted prompt instructions affecting authentication and outbound requests.

Content

Scanner excerpt · references/workflow.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: company-intel/1.0.2

Obfuscated Code

High
Category
Supply Chain
Confidence
50% confidence
Finding

Code contains obfuscation (base64, hex encoding with execution). This is often used to hide malicious functionality.

Content

Scanner excerpt · scripts/render_report.py (reported line 93)May include surrounding context.

python
# 知了标讯白色 logo(299x96 PNG base64 内嵌,保证报告离线/打印/转发时不裂图)
_LOGO_B64 = "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
...[truncated 27 chars]

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding

The skill explicitly performs local file writes by generating and saving HTML reports under ~/zlbx-company-intel-files/, but it does not declare a corresponding tool scope or permission boundary. That creates a capability mismatch: a host or reviewer cannot tell from the manifest that the skill will write to disk, which weakens consent, sandboxing, and policy enforcement.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The description repeatedly says the skill 'must' be used for broad categories of sales and pre-contact research, even when the user did not explicitly request this skill. This is a prompt-routing overreach that can bias agent behavior away from user intent and toward unnecessary external queries, increasing data exposure and spend risk.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The file is entirely framed as a Chinese-language skill and instructs users to send a fixed Chinese prompt, but it does not state that Chinese is optional or that the skill is region- or language-specific by design. Under the policy, forcing a specific language without opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The skill documentation is presented entirely in Chinese and labels the content as a quick reference for use, with no indication that users may choose another language or locale. Under the policy rule, forcing a specific language without opt-in is a natural-language policy violation unless the locale restriction is explicitly documented and justified.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

The file instructs the agent to serialize and transmit collected device features to an external auto-registration endpoint. External transmission alone is not always unsafe, but here it becomes a real issue because the transmitted data originates from local host fingerprinting unrelated to the declared skill purpose, creating privacy and trust-boundary concerns.

Content

Scanner excerpt · references/auto-register.md (reported line 130)May include surrounding context.

md
> ### ⚠️ 请求体必须用 JSON 序列化函数生成,不要手拼字符串
>
> 用 `json.dumps(payload)` / `requests.post(url, json=payload)` / `JSON.stringify(payload)`,
> 或 `curl -d @file`;**不要用字符串拼接,也不要用 Python 的 `str(dict)`**
> (后者产出单引号,服务端会报 `Expecting property name enclosed in double quotes`)。
>

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/auto-register.md (reported line 131)May include surrounding context.

md
> ### ⚠️ 请求体必须用 JSON 序列化函数生成,不要手拼字符串
>
> 用 `json.dumps(payload)` / `requests.post(url, json=payload)` / `JSON.stringify(payload)`,
> 或 `curl -d @file`;**不要用字符串拼接,也不要用 Python 的 `str(dict)`**
> (后者产出单引号,服务端会报 `Expecting property name enclosed in double quotes`)。
>
> 历史教训:曾有版本采集 `home_path`,Windows 的 `C:\Users\alice` 直接拼进 JSON 字符串时

Session Persistence

Medium
Category
Rogue Agent
Confidence
93% confidence
Finding

The document instructs writing the obtained API key into ~/.zlbx/config.json, merging with existing config, and immediately reusing it for subsequent requests. Persistent credential storage inside a skill is risky because it creates long-lived local secrets and cross-session effects unrelated to the core business-analysis task, and could expose or overwrite user environment state if the skill is misused or compromised.

Content

Scanner excerpt · references/auto-register.md (reported line 195)May include surrounding context.

text

注意事项:
- 目录不存在时先 `mkdir -p ~/.zlbx`
- 文件已存在时**合并而非覆盖**(保留用户可能的其他配置)
- `source: "auto"` 字段必须写入,**这是后续判断「是否输出自动登录链接」的关键依据**

Session Persistence

Medium
Category
Rogue Agent
Confidence
93% confidence
Finding

The document instructs writing the obtained API key into ~/.zlbx/config.json, merging with existing config, and immediately reusing it for subsequent requests. Persistent credential storage inside a skill is risky because it creates long-lived local secrets and cross-session effects unrelated to the core business-analysis task, and could expose or overwrite user environment state if the skill is misused or compromised.

Content

Scanner excerpt · references/auto-register.md (reported line 195)May include surrounding context.

text

注意事项:
- 目录不存在时先 `mkdir -p ~/.zlbx`
- 文件已存在时**合并而非覆盖**(保留用户可能的其他配置)
- `source: "auto"` 字段必须写入,**这是后续判断「是否输出自动登录链接」的关键依据**

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The document is written as a mandatory output template in Chinese and uses absolute instructions such as producing the report in this form, but it does not offer a language choice or state that Chinese is optional. That creates a natural-language locale policy concern because the skill appears to force a specific language regardless of user preference.

Content

No source excerpt is available for this finding.