ClawHub
Back to skill

Security audit

lovart skill

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent Lovart.ai API integration, but it can send design prompts to Lovart and may consume API credits.

Install this only if you intend to use Lovart.ai. Configure LOVART_API_KEY in a trusted environment, avoid sending confidential product plans, customer data, secrets, or unreleased brand material unless third-party processing by Lovart is acceptable, and consider confirming before broad design-generation requests that may use paid credits.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (9)

Vague Triggers

Medium
Confidence
82% confidence
Finding
The trigger phrases include very broad, generic requests such as 'AI 设计生成', '生成产品图片', and '创建营销素材', which can overlap with ordinary user intents unrelated to this specific integration. In an agent ecosystem, this can cause unintended activation and unnecessary routing of user prompts to an external API-backed skill, increasing privacy and cost exposure.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger phrases are broad enough to match many generic design and image-generation requests, which can cause the platform to invoke this skill unexpectedly. That increases the chance of misrouting user intent to an external API, leading to unintended data sharing, unnecessary API usage, or surprising behavior.

Natural-Language Policy Violations

Medium
Confidence
80% confidence
Finding
The manifest description and triggers are entirely Chinese-language, with no documented locale scoping or user-choice mechanism. This can cause incorrect activation behavior, exclusion of non-Chinese users, or unintended routing if the hosting system does not explicitly bind the skill to a Chinese locale.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The trigger phrases are broad enough to overlap with ordinary user requests about design or image generation, which can cause the skill to activate unexpectedly. In a skill that sends prompts and other user-supplied content to an external API, overbroad activation increases the chance of unintended data disclosure or unapproved third-party processing.

External Transmission

Medium
Category
Data Exfiltration
Content
- `template_id`: 模板 ID（如果适用）

### 第三步：调用 API
使用以下 curl 命令模板：

```bash
curl -X POST https://api.lovart.ai/v1/design/generate \
Confidence
91% confidence
Finding
The skill explicitly instructs transmitting user-provided design content to an external third-party service, which is a real data exposure boundary. Even though this is core to the skill's purpose, it can leak sensitive prompts, embedded personal data, or proprietary design information if users are not clearly informed and consent is not obtained.

External Transmission

Medium
Category
Data Exfiltration
Content
使用以下 curl 命令模板：

```bash
curl -X POST https://api.lovart.ai/v1/design/generate \
  -H "Authorization: Bearer $LOVART_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
89% confidence
Finding
The hardcoded external endpoint confirms that user content is sent off-platform to api.lovart.ai. In a skill whose main function is API integration this is expected, but it still represents a genuine external transmission risk because the prompt lacks strong disclosure, consent, and data-minimization guidance.

External Transmission

Medium
Category
Data Exfiltration
Content
**你的执行**:
```bash
curl -X POST https://api.lovart.ai/v1/design/generate \
  -H "Authorization: Bearer $LOVART_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
88% confidence
Finding
This example scenario operationalizes the same external transmission pattern by encouraging sending concrete user requests to the third-party API. Sample commands can normalize data sharing without warning users that their commercial or product information may be processed externally.

External Transmission

Medium
Category
Data Exfiltration
Content
**你的执行**:
```bash
curl -X POST https://api.lovart.ai/v1/design/generate \
  -H "Authorization: Bearer $LOVART_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
88% confidence
Finding
This social-media example again demonstrates sending prompt content to an external service, reinforcing the data-sharing behavior without sufficient privacy framing. While not malicious, repeated examples make the transmission pathway more likely to be used casually with sensitive marketing plans or unpublished campaign materials.

External Transmission

Medium
Category
Data Exfiltration
Content
**你的执行**:
```bash
curl -X POST https://api.lovart.ai/v1/design/generate \
  -H "Authorization: Bearer $LOVART_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
88% confidence
Finding
This logo-design example also sends potentially sensitive brand-concept information to a third-party endpoint. In context the transmission is functionally necessary, but the skill does not adequately warn about confidentiality, intellectual-property sensitivity, or external retention of submitted design inputs.

Static analysis

No suspicious patterns detected.