Back to skill

Security audit

每股收益·逐条挂源

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because scanners report external service and credential risks, privacy gaps, and behavior that does not clearly match its stated EPS extraction purpose; coverage was partial.

Consider not installing until the external destinations, token handling, and document transfers are clear and appropriate for your use. The verdict is suspicious, not malicious; the AIG scan reported no findings, but Skillspector coverage was partial (75%), with a tool-misuse analyzer degraded, so the evidence is incomplete.

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 (8)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied commands perform remote service discovery and authenticated tool enumeration rather than the declared financial-disclosure extraction task. Nothing in the chunk connects those services to EPS analysis, so the network access and token use are not established as supporting implementation details. The routing URL must already be supplied externally; printing catalog routing information does not assign ROUTING_URL.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

所提供代码没有读取财报或公告,也没有提取每股收益披露、保留原句或生成来源锚点。其全部可见行为是获取并打印外部 MCP 目录数据,未体现其与声明任务之间的支持关系,因此属于实质性用途及资源访问不匹配。代码未显示上传财务数据或其他本地信息。

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

The anonymous HTTPS catalog fetch is not itself sensitive exfiltration, but the following workflow sends CUE_API_KEY to a routing URL obtained from that remote catalog without specifying destination validation. A compromised catalog or substituted routing value could therefore direct the bearer credential to an attacker-controlled endpoint.

Content

Scanner excerpt · README.md (reported line 75)May include surrounding context.

发现先行,匿名取 live 目录,拿到 routing 后再连接发 tools/list,以现场返回为准,不硬编码工具名与工具数:

bash
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.get('routing') or d)"
# 拿到 routing 后,带 token 发 tools/list:
curl -sS -X POST "$ROUTING_URL" -H "Authorization: Bearer $CUE_API_KEY" \

The documented workflow issues HTTP requests despite stating that the skill makes no requests.

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

The README documents external catalog discovery and authenticated requests, then broadly states that the skill makes no requests. Although the commands are visible and are not automatically executed by the local validator, this contradictory assurance can mislead users or agents about the workflow's network and credential exposure.

Content

No source excerpt is available for this finding.

The attachment-parsing workflow lacks a warning that document contents are sent to a third-party service.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The workflow directs long attachments and scanned documents to an external parsing service without explaining that their contents leave the user's environment or requiring approval. Published annual reports reduce the usual sensitivity, but user-supplied attachments can contain annotations, confidential drafts, or other nonpublic material.

Content

No source excerpt is available for this finding.

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.

The attachment-parsing workflow lacks a privacy warning about sending user-provided documents to an external service.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The instructions initiate external parsing when evidence resides in an attachment, but do not require informed approval before transferring user-provided document contents. Public financial disclosures are generally less sensitive, yet the accepted input types do not exclude confidential or annotated documents.

Content

No source excerpt is available for this finding.

The validator permits self-calculated dilution adjustments without disclosure evidence, contrary to the manifest's disclosure-only requirement.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The quote checker exempts lines containing 〔自算〕 from disclosure evidence requirements, while the dilution checker explicitly skips those lines. Without a quotation, the numeric checker also skips them, allowing an unsupported self-calculated adjustment to pass despite the skill's disclosure-only contract and potentially giving misleading assurance about financial output.

Content

No source excerpt is available for this finding.