ClawHub
Back to skill

Security audit

AI图片生成

Security checks across malware telemetry and agentic risk

Overview

This is a straightforward SiliconFlow image-generation skill that sends prompts to the documented API using the user's API key.

Install only if you are comfortable sending image prompts to SiliconFlow and letting the skill use your SiliconFlow API key, which may consume paid quota. Avoid secrets, regulated data, or proprietary text in prompts, and make sure jq is installed before using the helper script.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (6)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The script transmits the full user-supplied prompt to a third-party image-generation service without any explicit consent, warning, or data-sensitivity check. In an agent context, users may enter secrets, personal data, or proprietary text into prompts, causing unintended external disclosure to the API provider.

External Transmission

Medium
Category
Data Exfiltration
Content
## Generate Image

```bash
curl -X POST "https://api.siliconflow.cn/v1/images/generations" \
  -H "Authorization: Bearer $SILICONFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
89% confidence
Finding
The hardcoded external endpoint confirms the skill depends on outbound network access to SiliconFlow. In context this is necessary functionality rather than malicious exfiltration, but it still creates a trust dependency on a third-party API and exposes any submitted content to an external domain.

External Transmission

Medium
Category
Data Exfiltration
Content
## Generate Image

```bash
curl -X POST "https://api.siliconflow.cn/v1/images/generations" \
  -H "Authorization: Bearer $SILICONFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
89% confidence
Finding
The hardcoded external endpoint confirms the skill depends on outbound network access to SiliconFlow. In context this is necessary functionality rather than malicious exfiltration, but it still creates a trust dependency on a third-party API and exposes any submitted content to an external domain.

External Transmission

Medium
Category
Data Exfiltration
Content
- `black-forest-labs/FLUX.1-dev` - FLUX dev version
- `Qwen/Qwen-Image` - Qwen image generation

> 注意：模型ID可能更新，可通过 `curl -s "https://api.siliconflow.cn/v1/models" -H "Authorization: Bearer $SILICONFLOW_API_KEY" | jq '.data[].id'` 查询最新列表

## Parameters
Confidence
87% confidence
Finding
The model-list query also sends the bearer token to the external SiliconFlow API, expanding the number of operations that transmit credentials externally. This is normal for service discovery, but it still represents outbound credential use and should be treated as a controlled external dependency.

External Transmission

Medium
Category
Data Exfiltration
Content
echo "Generating image with $MODEL..."
echo "Prompt: $PROMPT"

RESPONSE=$(curl -s -X POST "https://api.siliconflow.cn/v1/images/generations" \
    -H "Authorization: Bearer $API_KEY" \
    -H "Content-Type: application/json" \
    -d "{
Confidence
93% confidence
Finding
This code sends user-controlled content and an API credential to an external network endpoint, which creates a real data egress boundary. In this skill, external transmission is expected for functionality, but it still represents a vulnerability when done without transparent disclosure, minimization, or controls over what data can be exfiltrated.

External Transmission

Medium
Category
Data Exfiltration
Content
echo "Generating image with $MODEL..."
echo "Prompt: $PROMPT"

RESPONSE=$(curl -s -X POST "https://api.siliconflow.cn/v1/images/generations" \
    -H "Authorization: Bearer $API_KEY" \
    -H "Content-Type: application/json" \
    -d "{
Confidence
93% confidence
Finding
This code sends user-controlled content and an API credential to an external network endpoint, which creates a real data egress boundary. In this skill, external transmission is expected for functionality, but it still represents a vulnerability when done without transparent disclosure, minimization, or controls over what data can be exfiltrated.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.