Back to skill

Security audit

主要会计数据与财务指标·逐条挂源

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because scanners report credentialed requests to a configurable remote service and external document processing without clear disclosure for the stated financial-data task.

Consider the suspicious verdict and medium confidence before installing. Review whether you are comfortable with the reported credentialed requests to a configurable endpoint and external processing of report attachments; the scan was partial, with one script only partly inspected.

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
98% confidence
Finding

The supplied chunk performs service and tool discovery but contains no annual-report reading, source anchoring, metric registration, or financial reconciliation. Although tool discovery could support a larger workflow, neither the declared purpose nor this chunk establishes a connection between these external services and the accounting task. Its external resource access and credential use are therefore materially undeclared. The code does not establish that credentials are sent to cuecue.cn; they are sent to the separately configured MCP_ENDPOINT.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

所给代码仅执行远程服务发现和带凭据的工具枚举,没有读取年报、记录原文锚点或核算财务数据的行为。虽然工具发现可能作为某些系统的辅助步骤,但该代码未体现这些 MCP 资源与声明的年报处理任务有何联系,因此构成未声明能力及资源访问不一致。代码明确发送 API 凭据,但无法仅据此判定接收端是否恶意或凭据是否被窃取。

Content

No source excerpt is available for this finding.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 148)May include surrounding context.

md
| `scripts/fixtures/run_fixtures.sh` | 离线样例 runner:自测→好样→坏样三档,坏样注入必先断言落地;期望退码由自测末行的「不可判=N」现算,不写死 |

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

The catalog GET does not itself transmit financial documents or credentials, and remote tool discovery supports the stated workflow. However, the subsequent request sends CUE_API_KEY to a configurable MCP_ENDPOINT without a documented HTTPS or trusted-origin check; a substituted endpoint could receive the bearer credential.

Content

Scanner excerpt · README.md (reported line 112)May include surrounding context.

bash
  # ① 匿名取域清单与 routing
  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

The workflow sends report screenshots and scans to an external reader without a privacy warning.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The workflow directs screenshots and scans to an external parsing service without requiring disclosure or authorization before upload. Published annual reports are generally public, which reduces the risk, but user-supplied copies can contain confidential annotations 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.

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

The public catalog request is a legitimate discovery step and contains no document payload. The adjacent authenticated request nevertheless attaches CUE_API_KEY to a configurable endpoint without requiring trusted-origin or HTTPS validation, so manipulated routing or endpoint configuration could disclose the key.

Content

Scanner excerpt · SKILL.md (reported line 93)May include surrounding context.

bash
  # ① 匿名取域清单与 routing(不硬编码工具名与工具数)
  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:工具名、参数与分页以现场返回为准

The attachment-parsing workflow lacks a warning that user-provided documents may be sent to an external service.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

Attachment parsing is directed to an external service when source text is inside an attachment, without an explicit confidentiality check or upload-authorization requirement. Although the intended disclosures are usually public, user-provided attachments may include private annotations or unpublished information that would cross a trust boundary.

Content

No source excerpt is available for this finding.