Back to skill

Security audit

实控人变更与要约收购·逐条挂源

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence because it reports undeclared external MCP discovery and privacy concerns, plus a weakness in a financial-data check; analysis was partial.

The verdict is suspicious, not a claim that the author acted maliciously. Before installing, consider whether you trust the external parsing service and whether you would pass it URLs or paths that could reveal private information. The scan covered 12 of 13 components fully and one partially, so it cannot establish that the artifact is safe.

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 (15)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

所给代码实际执行的是MCP服务发现和工具枚举,而不是声明中的A股公告信息提取与登记。虽然工具发现可能作为实现细节,但此片段没有说明这些端点提供公告数据或登记能力,因此无法将其认定为与声明目的相符的辅助操作。此外,注释称完成握手并确认三个工具,但实际仅发送tools/list请求,没有执行初始化握手或校验工具数量。

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The supplied chunk performs only remote service discovery and tool enumeration. It contains no announcement processing, transaction classification, required-field extraction, or evidence anchoring. Although MCP discovery could support a larger implementation, this chunk does not establish that either endpoint provides resources relevant to the declared purpose. No data exfiltration or investment prediction is demonstrated.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The supplied chunk only discovers external tooling; it does not retrieve securities announcements, extract the required transaction fields, or attach source quotations and anchors. MCP discovery could support the declared workflow, but this chunk provides no link between the catalog or listed tools and that workflow. The assessment is limited to the supplied code: it does not establish malicious behavior or show any discovered tool being executed.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The supplied chunk performs only remote service discovery. It does not retrieve securities announcements, extract the required transaction fields, or attach quotations and anchors. MCP discovery could support the declared workflow, but this chunk establishes no such connection and does not implement that workflow. The mismatch is based on the demonstrated behavior, not merely on its use of network requests. No financial-data transmission or remote tool execution is shown.

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
bash scripts/fixtures/run_fixtures.sh # 离线样例三档(零网络零计费)

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.en.md (reported line 111)May include surrounding context.

curl -s https://cuecue.cn/api/mcp-catalog | head -c 500

2) complete the MCP handshake and confirm the three tools are present

curl -s -X POST "$MCP_ENDPOINT" -H 'Content-Type: application/json'
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

text

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 93)May include surrounding context.

curl -s https://cuecue.cn/api/mcp-catalog | head -c 500

2) 按 MCP 协议握手,确认三枚工具在场

curl -s -X POST "$MCP_ENDPOINT" -H 'Content-Type: application/json'
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

text

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 · SKILL.en.md (reported line 133)May include surrounding context.

curl -s https://cuecue.cn/api/mcp-catalog | head -c 500

2) complete the MCP handshake and confirm the three tools are present

curl -s -X POST "$MCP_ENDPOINT" -H 'Content-Type: application/json'
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

text

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 · SKILL.md (reported line 117)May include surrounding context.

curl -s https://cuecue.cn/api/mcp-catalog | head -c 500

2) 按 MCP 协议握手,确认三枚工具在场

curl -s -X POST "$MCP_ENDPOINT" -H 'Content-Type: application/json'
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

text

The parsing workflow sends filing URLs or local paths to an external service without a privacy warning.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The workflow directs the agent to submit filing URLs or local paths to an external parser without a privacy check or explicit authorization. Although published filings are usually public, private URLs can contain access tokens and local paths can reveal sensitive metadata; transmission of file contents is not established by this source.

Content

No source excerpt is available for this finding.

The parsing workflow sends report URLs or local file paths to an external service without a privacy warning.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The parsing instructions send report URLs or local paths across an external-service boundary without screening for private information or obtaining specific authorization. Public-disclosure use reduces typical exposure, but user-supplied URLs may contain secrets and paths may expose confidential metadata; file-content upload is not demonstrated here.

Content

No source excerpt is available for this finding.

The workflow sends user-supplied filing URLs or local paths to an external parsing service without a privacy warning.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

The operational instructions make external parsing the normal workflow without requiring a privacy assessment or authorization for user-supplied inputs. Published securities reports are generally public, but signed URLs and local-path metadata can still be sensitive; the text does not establish that local file contents are uploaded.

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.

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 parsing workflow sends user-supplied report URLs or local paths to an external service without a privacy warning.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The skill directs submission of user-provided report URLs or local paths to an external service without a privacy gate or explicit authorization. While public announcements reduce ordinary sensitivity, private URLs may expose access tokens and local paths may disclose internal metadata; actual file-content upload is not established.

Content

No source excerpt is available for this finding.

The premium-rate check accepts an unrelated item-count declaration instead of the required pricing benchmark.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The benchmark check reports a missing basis only when both DECL.search(q) and BASIS.search(q) fail, so an unrelated item-count declaration inside the premium quotation suppresses the required warning. A crafted card can therefore pass this check with an unsupported premium benchmark, undermining a stated financial-data integrity safeguard.

Content

No source excerpt is available for this finding.