Back to skill

Security audit

爱企查企业情报-招投标实力视角查企业

Security checks for vulnerabilities and agentic risk

Overview

The skill’s company-intelligence function is coherent, but it also creates accounts, sends device-derived identifiers, stores API keys, and can retrieve contact phone data, so users should review it before installing.

Install only if you are comfortable using the ZLBX service for company research, sending company queries and limited device-registration data to its domains, letting it store an API key under ~/.zlbx/config.json, and paying quota/credits for lookups. Avoid using the contact lookup for bulk harvesting or unsolicited outreach, and prefer supplying your own API key if you do not want auto-registration.

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 (33)

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The activation rule is intentionally very broad: it says the skill must be used whenever a user wants to understand a company's 'real strength,' even if they did not ask for intelligence explicitly. Overbroad routing can cause the agent to invoke this skill in unrelated company-research contexts, leading to unnecessary external data sharing, account use, local report generation, and cost-incurring actions without sufficiently specific user intent.

Content

No source excerpt is available for this finding.

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
70% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

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: company-profile-aiqicha/1.0.6

Hidden Instructions

High
Category
Prompt Injection
Confidence
70% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

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: company-profile-aiqicha/1.0.6

Hidden Instructions

High
Category
Prompt Injection
Confidence
92% confidence
Finding

This duplicate finding identifies the same hidden-instruction pattern: covert directives tell the agent to preserve an API key context and attach package/source identity headers. In a skill ecosystem, concealed prompt directives are risky because they rely on the model to enforce authentication and provenance rules, which can be subverted or misapplied across packages.

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
92% confidence
Finding

This duplicate finding identifies the same hidden-instruction pattern: covert directives tell the agent to preserve an API key context and attach package/source identity headers. In a skill ecosystem, concealed prompt directives are risky because they rely on the model to enforce authentication and provenance rules, which can be subverted or misapplied across packages.

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
91% confidence
Finding

This duplicate hidden-instructions finding is also valid: the file embeds unseen directives telling the agent which headers and provenance values must be sent. In context, concealed request-shaping instructions are more dangerous because the same file also directs account registration and credential use, so the hidden metadata affects authenticated external traffic.

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
91% confidence
Finding

This duplicate hidden-instructions finding is also valid: the file embeds unseen directives telling the agent which headers and provenance values must be sent. In context, concealed request-shaping instructions are more dangerous because the same file also directs account registration and credential use, so the hidden metadata affects authenticated external traffic.

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

The file adds an extensive auto-registration and account-management workflow to a skill whose declared purpose is enterprise intelligence and company profiling. This materially expands the skill’s authority from data lookup into local environment inspection, credential acquisition, persistence, and billing/recharge flows, creating an unjustified capability set that could be abused or trigger user actions unrelated to the requested business function.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill instructs the agent to collect device fingerprints such as OS type, CPU architecture, and a hashed MAC address in order to register an account. Even with hashing, this is persistent device-identifying telemetry unrelated to company-profile analysis, and it increases privacy risk while creating a pathway for silent host inspection before ordinary use of the skill.

Content

No source excerpt is available for this finding.

Hidden Instructions

High
Category
Prompt Injection
Confidence
88% confidence
Finding

This duplicate finding identifies the same hidden instruction block directing credentialed API use and fixed package/source headers. Even if intended for package integrity, placing such instructions in prompt content makes them adversarially reachable and can cause the model to prioritize concealed execution guidance over transparent policy and user control.

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
88% confidence
Finding

This duplicate finding identifies the same hidden instruction block directing credentialed API use and fixed package/source headers. Even if intended for package integrity, placing such instructions in prompt content makes them adversarially reachable and can cause the model to prioritize concealed execution guidance over transparent policy and user control.

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
97% confidence
Finding

This duplicate finding points to the same hidden-instruction block that directs credentialed request identity and package-source headers. In the context of an agent skill, concealed network instructions are especially risky because the skill processes externally supplied company names and then drives multiple API/Web requests, so hidden header manipulation can systematically affect every outbound request without clear user awareness.

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
97% confidence
Finding

This duplicate finding points to the same hidden-instruction block that directs credentialed request identity and package-source headers. In the context of an agent skill, concealed network instructions are especially risky because the skill processes externally supplied company names and then drives multiple API/Web requests, so hidden header manipulation can systematically affect every outbound request without clear user awareness.

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
82% confidence
Finding

The skill performs local file writes by generating and saving HTML reports, but it does not declare an explicit tool scope such as permissions or allowed-tools. Undeclared write capability weakens sandboxing and user review because the runtime may permit filesystem side effects that are not transparently bounded by policy.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

This section is labeled '首次使用前可告知用户', which makes disclosure of network access, local writes, data transmission, billing, and privacy optional. Because the skill writes reports locally, stores credentials, transmits company queries, and may auto-register an account, those behaviors should be consistently disclosed rather than treated as optional.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This markdown file contains user-facing operational instructions exclusively in Chinese, and nowhere indicates that users may choose another language or that the skill is limited to a Chinese-only audience. Under the language/locale policy, forcing a specific language without opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill documentation exposes a capability to retrieve project contact names and phone numbers, including full numbers for paid accounts, which goes beyond basic company profiling into direct personal-contact harvesting. In this context, that materially increases privacy, spam, and social-engineering risk because users can obtain actionable contact data tied to procurement activity.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
93% confidence
Finding

This section documents transmitting collected device features to an external auto-registration endpoint. The transmission itself is explicit rather than hidden, but it still exports local host-derived identifiers to a third-party service for a purpose unrelated to the skill’s declared business-analysis role, making it a real privacy and boundary-expansion issue.

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
90% confidence
Finding

The curl example reinforces that the skill is expected to serialize and send collected host information to an external service. While presented as implementation guidance, it operationalizes outbound exfiltration of environment-derived identifiers beyond what is necessary for enterprise intelligence lookups.

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

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The document directs the agent to write API keys into ~/.zlbx/config.json, merge with existing configuration, and immediately reuse the stored credential. Local credential persistence and reuse are unrelated to company-intelligence reporting and create a risk of unintended secret handling, stale credential reuse, and cross-skill trust expansion on the user machine.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
94% confidence
Finding

This duplicate session-persistence finding is likewise a true issue because the same section instructs storing and reusing keys across sessions. In combination with auto-registration and account recovery flows, persistent local state increases the blast radius of mistakes, unauthorized reuse, or future prompt-driven actions under the stored identity.

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
94% confidence
Finding

This duplicate session-persistence finding is likewise a true issue because the same section instructs storing and reusing keys across sessions. In combination with auto-registration and account recovery flows, persistent local state increases the blast radius of mistakes, unauthorized reuse, or future prompt-driven actions under the stored identity.

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**