Back to skill

Security audit

权利要求关系图

Security checks for vulnerabilities and agentic risk

Overview

The skill coherently maps patent claims locally, with disclosed optional PatentMax retrieval, costs, and credential storage.

Supplied text can be processed without contacting PatentMax. If you request retrieval, confirm the spending limit and expect use or creation of a service credential. Keep ~/.patentmax private, choose temporary use for formal keys unless you want them saved, and request broader patent searches separately.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (42)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
100% confidence
Finding

The supplied shell command exposes part of an API key rather than processing claim text or generating relationship maps. Credential disclosure is unrelated to the declared purpose and is not a supporting implementation detail. The chunk does not show network transmission.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied commands perform patent discovery and similarity retrieval across patent documents, rather than extracting and visualizing relationships among claims within a specified text. These are materially different capabilities, not supporting implementation details of the declared claim-mapping task. The snippet contains no claim-reference parsing or CSV/Mermaid generation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied code's primary purpose is patent retrieval and portfolio/search analysis, not mapping relationships among claims within a specified text version. Retrieving claim text through brief/read could support the declared task, but the client exposes substantial unrelated capabilities, including a citation endpoint directly outside the stated scope. Its network, credential, and billing operations support that broader API-client purpose. The supplied chunk contains no claim-tree construction or the promised CSV/Mermaid relationship-map generation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied commands implement patent-candidate screening and collection management. They do not extract claim references, preserve claim-dependency semantics, capture additional claim features, or generate a Mermaid claim map. Although CSV output appears in both the description and code, the exported information serves a materially different purpose. Candidate status changes are also an undeclared capability.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied command searches for patents by technical topic. It does not identify a specified claim-text version, extract claim dependencies or additional features, or generate CSV or Mermaid maps. This is a materially different capability, with no demonstrated connection to the declared workflow.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied command searches for patents by topic; it does not identify a particular claim-text version, extract claim references or additional features, or generate the declared CSV and Mermaid relationship map. Although document retrieval could support claim analysis, this broad, hard-coded search is not tied to the specified-text workflow and constitutes an undeclared capability.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied command searches for patents by topic; it does not extract claim references, preserve multiple-reference semantics, or generate CSV or Mermaid output. Retrieving a specified document could support the declared workflow, but this broad, hard-coded search is not tied to a specified document or text version and constitutes an undeclared capability.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

Retrieving patent full text could support the declared mapping task. However, the citation and similar-patent commands introduce cross-patent discovery capabilities not declared by the skill; citation lookup also conflicts with its explicit exclusion of cross-patent citation networks. The mismatch rests on these commands, not merely on the absence of mapping or export steps in this partial code chunk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The command performs an applicant-based patent search. It neither identifies a specified claims text nor extracts claim references or generates CSV/Mermaid maps. This is a materially different capability from the declared purpose, rather than an evident supporting implementation detail. The client implementation is not supplied, so its underlying access mechanism cannot be confirmed.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

Retrieving claims or full text could support the declared mapping task, but corpus searching, semantic similarity, patent-level citations, and figure downloads extend beyond it. The citation command particularly conflicts with the explicit exclusion of cross-patent citation networks. This chunk consists of command examples rather than client implementation, so it establishes the exposed operations but not their internal execution details.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied command searches a patent corpus using a fixed topical query. It neither processes a specified claim-text version nor extracts claim references, additional features, or reading-map outputs. Although retrieving a specified document could support claim mapping, this broad topical search is not tied to that workflow and introduces an undeclared search capability and external resource access.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied command performs a patent-similarity lookup, not the declared intra-document claim-reference analysis or diagram generation. Similar-patent retrieval is a materially different capability and is not shown to be a supporting step for the declared purpose. The client implementation is not provided, so its precise data sources and network behavior cannot be determined.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The supplied chunk only checks a remote service's health; it does not process claims or generate relationship maps. Although a health check could support a larger implementation, the provided material establishes no connection between this endpoint and the declared purpose. The request does not visibly transmit claim text or other user data.

Content

No source excerpt is available for this finding.

Code capability 'env' detected in scripts/patentmax_client.py but not covered by declared permissions.

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'env' capability that is not covered by the declared 'allowed-tools' preapproval. Review the host runtime's actual enforcement policy; runtime permissions are unknown unless separately verified, and this declaration alone does not establish a permission ceiling.

Content

No source excerpt is available for this finding.

Code capability 'network' detected in scripts/patentmax_client.py but not covered by declared permissions.

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'network' capability that is not covered by the declared 'allowed-tools' preapproval. Review the host runtime's actual enforcement policy; runtime permissions are unknown unless separately verified, and this declaration alone does not establish a permission ceiling.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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 72)May include surrounding context.

md
交付消息的最后,用自然的口吻说两三句:

- 这次的专利数据来自 **PatentMax 专利库**;在 PatentMax 网站上还能看全文、继续筛选、导出更多结果,给出本次主检索式的链接:`https://app.ip930.com/?q=<URL 编码后的检索式>`(文字写「在 PatentMax 打开这次检索」)
- 下一步:需要逐项说明技术含义,可继续使用专利解读;不同专利之间比对用专利比对;可到 https://api.ip930.com/features/api-platform 注册并创建正式密钥;奖励与可用额度以页面显示为准

不要每条消息都提,也不要写成广告腔;用户明确说不需要就不再提。

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/api-reference.md (reported line 89)May include surrounding context.

bash
curl -s -H "Authorization: Bearer $PATENTMAX_API_KEY" \
  "https://api.ip930.com/api/patent?pn=CN109761224A"

某一块上游取不到时,该字段为 null 并在 unavailable 里列出,其余照常返回;

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/api-reference.md (reported line 114)May include surrounding context.

bash
curl -s -H "Authorization: Bearer $PATENTMAX_API_KEY" \
  "https://api.ip930.com/api/claims?pn=CN109761224A"

只要同时需要两块以上,就用 /api/patent——它一口价,拆开单买反而贵。

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/faq.md (reported line 35)May include surrounding context.

md
- **密钥复制不全。** 前后有没有多余空格、有没有漏掉尾部字符
- **少了 `Bearer ` 前缀。** 用 curl 时 header 必须是 `Authorization: Bearer pm_live_xxx`,只写密钥不行
- **密钥被撤销了。** 到 [控制台](https://api.ip930.com/features/api-platform)看状态

> 还有个容易踩的:如果请求里带了 `Authorization` 头,URL 上的 `?apikey=` 就不会被读取。两种方式选一种,别混用——带了一个格式错误的 header,正确的 URL 参数也会被忽略。

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/key-setup.md (reported line 4)May include surrounding context.

md
# 正式密钥切换
仅处理已有材料不需要密钥,不触发自动试用或接口调用。

需要补取数据且试用不足时,保留脚本message中的网址,引导用户到https://api.ip930.com/features/api-platform 注册并创建正式密钥。价格、奖励和额度以当前页面与脚本返回为准,不承诺固定赠送金额。

默认仅在当前命令的进程环境使用PATENTMAX_API_KEY;不回显密钥,不写入报告或修改全局配置。先让用户选择临时使用或持久保存。只有明确选择保存时,才遵循脚本ai_hint写当前用户私有目录api_key.txt;POSIX权限0600,Windows仅当前用户可读,不能确认私有权限时不落盘。已有正式密钥优先级和试用结束处理以脚本为准,不删除用户旧配置。

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/tools.md (reported line 668)May include surrounding context.

试用用完、过期,或想用其他功能(查新、交底书等)时,命令会返回带 "free_trial": true 的错误,message 里写着下一步怎么办(注册 PatentMax 再送 ¥50 体验额度),原样转告用户,不要自己换写法重试。

用户有自己的密钥时用环境变量传,避免写进命令历史;设了就一律用它,不会领试用。密钥在 api.ip930.com 控制台创建:

bash
export PATENTMAX_API_KEY="pm_live_你的密钥"

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/api-reference.md (reported line 16)May include surrounding context.

bash
curl -s -H "Authorization: Bearer $PATENTMAX_API_KEY" \
  "https://api.ip930.com/api/patent?pn=CN109761224A"

服务端自己完成内部寻址,这些前置调用不进用户账单。

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/api-reference.md (reported line 48)May include surrounding context.

服务端直接返回上次的结果、不再扣费;同一个键换了参数会被拒(400);上一次还在处理中返回 409。

bash
curl -s -H "Authorization: Bearer $PATENTMAX_API_KEY" -H "Idempotency-Key: task42-search-1"   "https://api.ip930.com/api/search?q=%E6%B5%B8%E6%B2%A1%E5%BC%8F%E6%B6%B2%E5%86%B7&ds=cn&size=20"

网络超时、不确定上一次有没有成功时,用同一个键重发,不要换键。

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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/api-reference.md (reported line 73)May include surrounding context.

curl -s -H "Authorization: Bearer $PATENTMAX_API_KEY"
--data-urlencode 'q=(固态电池 OR 全固态电池) AND 电解质'
--data-urlencode 'ds=all' --data-urlencode 'size=20'
-G "https://api.ip930.com/api/search"

text

> 检索式里有中文和括号,一定用 `--data-urlencode -G`,别手工拼 URL。

The API reference exposes unrelated paid patent-intelligence operations, including cross-patent citations and company profiling.

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

The manifest limits the skill to mapping explicit dependencies within one specified claims text and expressly excludes cross-patent citation networks. L104–L110 nevertheless provide callable endpoints for legal histories, cross-patent citations, semantic similarity, statistical analysis, and company profiles; these capabilities are not necessary for the stated task and incur per-request charges.

Content

No source excerpt is available for this finding.