Back to skill

Security audit

clawtip-invention-patent

Security checks for vulnerabilities and agentic risk

Overview

This paid patent-drafting skill is broadly coherent, but it sends patent topics and payment credentials to a remote service over plaintext HTTP, so it belongs in Review before installation.

Install only if you are comfortable sending invention topics, order numbers, and payment credentials to the publisher's remote service. Do not use it for trade-secret-level or pre-filing confidential material unless the service is moved to HTTPS and provides clear privacy, retention, and payment-security terms.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/service.py:13
Finding

Payment Credentials and Patent Information Transmitted Over Plaintext HTTP

Content
View full analysis
None: payload = json.dumps( { "question": question, "orderNo": order_no, "credential": credential, }, ensure_ascii=False, ).encode("utf-8") req = urllib.request.Request( GET_RESULT_URL, data=payload, headers={"Content-Type": "application/json"}, method="POST", ) try: with urllib.request.urlopen(req, timeout=120) as resp: body = json.loads(resp.read().decode("utf-8")) ``` Order creation also uses the same plaintext HTTP service and sends the user's invention topic: ```python DEFAULT_BASE = os.environ.get("SKILL_SERVER_BASE_URL", "http://api.clawtip.top").rstrip("/") CREATE_ORDER_URL = f"{DEFAULT_BASE}/api/clawtip/createOrder" def create_order(question: str) -> tuple[str, int | str, str, str]: payload = json.dumps( {"question": question, "clientIp": ""}, ensure_ascii=False, ).encode("utf-8") req = urllib.request.Request( CREATE_ORDER_URL, data=payload, headers={"Content-Type": "application/json"}, method="POST", ) try: with urllib.request.urlopen(req, timeout=60) as resp: body = json.loads(resp.read().decode("utf-8")) ``` The documented workf ...[truncated 3002 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • 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 (14)

Tainted flow: 'req' from os.environ.get (line 22, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
94% confidence
Finding

The request target is derived from an environment-controlled base URL and the default uses plain HTTP, so the script can transmit user-supplied invention details and payment-order metadata to an attacker-controlled or intercepted endpoint. In this skill context, the data is especially sensitive because invention descriptions may contain unpublished patentable material, and loss of confidentiality can materially harm the user's IP position.

Content

Scanner excerpt · scripts/create_order.py (reported line 29)May include surrounding context.

python
method="POST",
    )
    try:
        with urllib.request.urlopen(req, timeout=60) as resp:
            body = json.loads(resp.read().decode("utf-8"))
    except urllib.error.URLError as e:
        raise RuntimeError(f"网络请求异常,请确认服务地址 {CREATE_ORDER_URL} 可达: {e}") from e

Tainted flow: 'req' from os.environ.get (line 26, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
97% confidence
Finding

The request destination is derived from the SKILL_SERVER_BASE_URL environment variable and sent directly via urllib.request.urlopen without validation or enforcement of a trusted HTTPS origin. Because this request carries the payment credential, a malicious or compromised runtime environment can redirect it to an attacker-controlled server and exfiltrate payment-related secrets or user data.

Content

Scanner excerpt · scripts/service.py (reported line 33)May include surrounding context.

python
method="POST",
    )
    try:
        with urllib.request.urlopen(req, timeout=120) as resp:
            body = json.loads(resp.read().decode("utf-8"))
    except urllib.error.URLError as e:
        print("PAY_STATUS: ERROR")

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The documentation promises multi-turn Chinese patent drafting, but the described implementation mainly performs CLI parameter handling and external payment/status verification. In a legal-tech context involving potentially confidential invention disclosures, overstating local functionality while depending on opaque remote behavior increases the chance of unintended data exposure and user deception.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The documentation promises multi-turn Chinese patent drafting, but the described implementation mainly performs CLI parameter handling and external payment/status verification. In a legal-tech context involving potentially confidential invention disclosures, overstating local functionality while depending on opaque remote behavior increases the chance of unintended data exposure and user deception.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 164)May include surrounding context.

md
- `SKILL.md` 中第二阶段 JSON 已按需填写 `skill_name`、`description`、`resource_url`

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The code defaults to http://api.clawtip.top and includes the payment credential in the POST body, which exposes the credential and associated request contents to interception or modification in transit. In a payment-gated patent drafting skill, this is especially dangerous because it handles paid access tokens and potentially sensitive invention details, increasing both financial and confidentiality risk.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

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.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The description explicitly instructs the skill to use Chinese for all interaction, which is a language-policy constraint. The file does not offer the user any language choice or opt-in, and the mandate is repeated later in the document.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
89% confidence
Finding

This documents outbound transmission to a third-party HTTP endpoint for order creation and fulfillment. External transmission is not inherently unsafe, but in this context it involves sensitive invention and payment workflow data, and the use of plain http:// rather than https:// raises confidentiality and integrity concerns because traffic may be intercepted or modified in transit.

Content

Scanner excerpt · SKILL.md (reported line 31)May include surrounding context.

md
**计费:** 每次有效会话 **2.00 元人民币**(OpenClaw 中 `AMOUNT=200`,单位为分)。

**本技能服务地址:** `http://api.clawtip.top/api`(创建订单、履约校验均指向该地址下的接口;若需覆盖,可设置环境变量 `SKILL_SERVER_BASE_URL`)。

---

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill sends payment and invention-topic data to an external service but lacks a clear privacy warning before transmission. Because patent topics can contain commercially sensitive, pre-filing invention details, failing to warn users meaningfully increases the risk of confidential information being disclosed to a third party without informed consent.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
89% confidence
Finding

The payment workflow example includes a resource_url pointing to an external HTTP service, confirming that payment-related and topic-related metadata are intended for third-party transmission. In a paid legal-tech skill, sending such data over insecure transport can expose confidential business information and payment context to interception or tampering.

Content

Scanner excerpt · SKILL.md (reported line 114)May include surrounding context.

"encrypted_data": "<ENCRYPTED_DATA>", "description": "中国发明专利撰写引导(单次)", "skill_name": "发明专利撰写助手", "resource_url": "http://api.clawtip.top/api" }

text

External Transmission

Medium
Category
Data Exfiltration
Confidence
87% confidence
Finding

The repeated publication of the external service endpoint reinforces that the skill depends on a remote third-party server for core behavior. The risk is elevated by the skill's subject matter: users may share patentable invention details before filing, so any insecure or undisclosed transmission can harm confidentiality and competitive position.

Content

Scanner excerpt · SKILL.md (reported line 172)May include surrounding context.

md
## 联系与扩展

服务端实现可参考仓库内 `api/clawtip` 与《技能开发指南》。线上服务地址:**`http://api.clawtip.top/api`**。若需调整单价,请同步修改服务端配置 `order_amount_fen` 与本说明中的用户展示文案。

---

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The script sends the user's invention question to a remote API as part of order creation without any visible warning, consent prompt, or confidentiality disclosure. For a patent-drafting assistant, this is particularly dangerous because users may provide novel technical details before filing, creating a serious confidentiality and trade-secret exposure if the service is compromised, logged, or untrusted.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

User-facing strings including the module description, argument description, and error/help text are written only in Chinese. This can violate language/locale policy when no user opt-in or documented justification is provided for restricting the interface to a specific language.

Content

No source excerpt is available for this finding.