Back to skill

Security audit

金融工具风险敞口·到期分档与敏感性重算

Security checks for vulnerabilities and agentic risk

Overview

The skill is marked suspicious because scanners report that its documented zero-network assurance conflicts with optional external catalog and document-processing workflows, and the scan was only 75% complete.

Consider the skill suspicious and review its network and document-sharing behavior before installing. The scanners report conflicting network assurances and possible external document processing; they do not establish malicious intent. Coverage was partial (75%), including a degraded tool-misuse analysis, so the results do not provide a complete safety assessment.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (5)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The source makes categorical zero-network and no-upload assurances while also enabling external retrieval and document parsing when an API key is present. Those integrations are related to financial extraction and explicitly mentioned, so the finding overstates their unrelatedness, but the conflicting privacy contract can still lead to unintended external processing.

Content

No source excerpt is available for this finding.

The workflow sends screenshots and scanned documents to an external reader service without a clear data-sharing warning.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The workflow directs screenshots and scanned documents to an external reader without an explicit disclosure and consent step for document transmission. Annual reports are often public, which reduces typical exposure, but user-supplied files can contain annotations or nonpublic financial information, especially when users rely on the zero-network claim.

Content

No source excerpt is available for this finding.

The documented agent setup makes external network requests despite the README's explicit zero-network claim.

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

The README promises zero-network operation but also instructs agents to perform external catalog discovery and authenticated tool enumeration. Although the integrations are documented, the unqualified offline assurance can cause users or agents to cross a network trust boundary they believed would remain closed.

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 · README.md (reported line 110)May include surrounding context.

bash
  # ① 匿名取活体清单
  curl -sS https://cuecue.cn/api/mcp-catalog -o /tmp/mcp-catalog.json
  python3 -c "import json;d=json.load(open('/tmp/mcp-catalog.json'));print(d)"

  # ② 带凭据握手取工具表

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read, file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.