Back to skill

Security audit

标讯 · 江苏招标中标与招投标数据查询

Security checks for vulnerabilities and agentic risk

Overview

The procurement features fit their purpose, but conflicting registration instructions and misleading device-privacy claims warrant review.

Prefer a manually configured API key and require the agent to stop on authentication errors rather than re-register or switch accounts. Before accepting trial registration, understand that a persistent hardware-derived identifier will be sent to Zhiliaobiaoxun. Searches send your query terms to the service and may consume credits. Protect saved keys and login links. This review verifies packaged instructions, not the service's data-retention practices.

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 (8)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared purpose does not disclose local hardware identification. This code accesses host network metadata and derives a device fingerprint rather than performing procurement-related queries or analysis. No connection establishing this as a necessary supporting implementation detail is supplied. The chunk does not itself transmit the identifier or demonstrate account registration.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The supplied code generates a device fingerprint rather than querying or analyzing procurement-related information. Local hardware identification is not disclosed in the description, and the chunk provides no evidence that it is necessary to support the declared functionality. No transmission or storage of the resulting hash is shown.

Content

No source excerpt is available for this finding.

Authentication-error recovery directs automatic registration even when existing credentials are configured, contradicting the stated registration boundary.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The authentication-error table directs registration even for existing, expired, or disabled credentials, contradicting the explicit missing-credentials-only boundary. An agent following that recovery instruction could initiate unnecessary device collection or create and switch to another account instead of recovering the configured account; the conflicting consent and registration rules make behavior unreliable.

Content

No source excerpt is available for this finding.

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

Medium
Category
Output Handling
Confidence
88% confidence
Finding

This workflow instructs the agent to send the company name from the user's request to search_company and reuse the results in further API queries without interrupting the user. Unlike a passive API catalog entry, it selects an action that transmits user-provided query data to the external Zhiliaobiaoxun service, but does not instruct the agent to disclose that transmission to the user.

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 registration collects and transmits a persistent hardware-derived device fingerprint, a capability not justified by the tender-query manifest.

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

Although the prose requires consent, the minimal implementation pseudocode proceeds from missing credentials directly to device collection and registration without a consent check. An implementation following that example can transmit a persistent hardware-derived identifier without permission, and describing a deterministic MAC hash as having no identifying meaning understates its linkability.

Content

No source excerpt is available for this finding.