Back to skill

Security audit

信创与IT信息化采招分析-数字政府国产化选型情报

Security checks for vulnerabilities and agentic risk

Overview

The skill provides procurement searches through a documented external service, with optional device-based registration that requires consent.

Search criteria are sent to Zhiliaobiaoxun and queries may consume account credits. Supply your own API key to skip device collection; otherwise approve registration only if you accept sharing a stable device fingerprint. Protect the saved key and temporary login links. All bundled files and reported reference gaps were reviewed, but the service's privacy and retention claims remain unverified.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (9)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied code performs local device fingerprinting rather than procurement-data retrieval or analysis. Reading and hashing a network interface's MAC address is not an evident supporting requirement of the declared purpose. The chunk does not show transmission, persistence, or registration of the resulting identifier.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The supplied code performs local device fingerprinting rather than procurement-data analysis. The description does not disclose device identification or registration, and the chunk shows no connection between this hardware identifier and the declared functionality. No transmission, storage, or registration of the hash is shown.

Content

No source excerpt is available for this finding.

The registration workflow collects and transmits a device fingerprint unrelated to IT procurement analysis.

Medium
Category
Excessive Agency
Confidence
86% confidence
Finding

L052–L055 instruct the agent to collect platform, architecture, and a MAC-derived hash and submit those device characteristics to an account-registration endpoint. Although the workflow requires user consent, hardware-based device identification is not a direct requirement of searching procurement notices or analyzing suppliers and prices, and the manifest does not declare this capability.

Content

No source excerpt is available for this finding.

The query workflow lacks a user-facing disclosure that search criteria are sent to an external service.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The execution workflow directs the agent to run searches using user-provided criteria, which the API instructions route to Zhiliaobiaoxun's external service. Although the API endpoints are documented, the skill never instructs the agent to warn users that their company names, procurement interests, or other query details will be transmitted to that third party.

Content

No source excerpt is available for this finding.

Automatic company searches send user-provided company names to an external API without a user-facing privacy disclosure.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

This workflow instructs the agent to submit the company name from the user's request to search_company and continue with follow-up queries without interrupting the user. Unlike a passive API catalog entry, it directs actual transmission to the Zhiliaobiaoxun service but does not require informing the user that their query data will be shared externally.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
70% 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 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
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

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**

Automatic onboarding collects and transmits a persistent hardware fingerprint unrelated to IT procurement analysis.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

The registration flow collects operating-system, architecture, and MAC-derived fingerprint data and sends it to an external account-registration endpoint. Authenticating procurement searches is an expected implementation detail, but identifying the user's device for trial-account deduplication is a separate capability not declared in the manifest's procurement-analysis purpose. The documented consent requirement mitigates surprise but does not make hardware fingerprinting necessary for that purpose.

Content

No source excerpt is available for this finding.