ClawHub
Back to skill

Security audit

高德打车Skill

Security checks for vulnerabilities and agentic risk

Overview

The scanners classify this ride-hailing skill as benign, with medium confidence: they found no AIG issues, while the other scanner flagged ambiguous invocation examples, limited privacy disclosure, and a connectivity-check command; review coverage was partial.

The verdict is benign, but confidence is medium and the skillspector scan is marked partial: it inspected all 11 listed components, yet some runtime-related references could not be resolved. Before installing, consider whether you are comfortable sharing trip locations, timing, and preferences with Amap, and whether the ambiguous examples fit your intended use.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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
Findings (3)

Invocation examples include requests that do not establish ride-hailing intent.

Medium
Category
Trigger Abuse
Confidence
90% confidence
Finding

The examples include “查询订单” (“query orders”), which could refer to shopping or other services, and “明天早上8点从家到公司” (“tomorrow at 8 a.m. from home to work”), which could request directions rather than a taxi. Although the README describes an Amap ride-hailing skill, it does not restrict these ambiguous phrases to an already established ride-hailing conversation.

Content

No source excerpt is available for this finding.

The planning workflow sends sensitive trip information to an external service without requiring a user-facing privacy disclosure.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The workflow instructs the agent to send collected trip parameters to the Amap ride-hailing service, including origin and destination addresses that may be populated from saved home or workplace preferences. It does not instruct the agent to tell users that these details will be shared externally; confirmation of a booking occurs only after the initial planning call.

Content

No source excerpt is available for this finding.

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/errors.md (reported line 34)May include surrounding context.

现象：调用返回连接错误；工具列表里没有打车工具时按 auth.md「失败分类」处理。$MCP_URL 的配置见 auth.md「命令行排查」。连通性探测：

bash
curl -X POST "$MCP_URL" -H "Content-Type: application/json" -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"client","version":"1.0"}}}'