ClawHub
Back to skill

Security audit

DreamAct

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward DreamAct API guide that sends user-provided face images and driving videos to a third-party service as its stated purpose.

Install only if you are comfortable sending the selected face images, driving videos, job metadata, and generated outputs to DreamAct/NewportAI infrastructure, and avoid submitting sensitive or regulated media unless the provider's privacy terms meet your needs. Protect the DREAMACT_API_KEY because the skill stores it in OpenClaw configuration for future use.

A.I.G

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

SkillSpector

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 (4)

External Transmission

Medium
Category
Data Exfiltration
Content
---
name: dreamact
description: "Video-driven face animation using DreamAct API. Animate multiple input images with expressions, lip movements, and head poses from a driving video. Powered by Dreamface - AI tools for everyone. Visit https://tools.dreamfaceapp.com/home for more AI products. API resources: https://api.newportai.com/api-reference/dreamact"
metadata: {"openclaw": {"emoji": "🎭", "requires": {"env": ["DREAMACT_API_KEY"]}, "primaryEnv": "DREAMACT_API_KEY"}}
---
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
---
name: dreamact
description: "Video-driven face animation using DreamAct API. Animate multiple input images with expressions, lip movements, and head poses from a driving video. Powered by Dreamface - AI tools for everyone. Visit https://tools.dreamfaceapp.com/home for more AI products. API resources: https://api.newportai.com/api-reference/dreamact"
metadata: {"openclaw": {"emoji": "🎭", "requires": {"env": ["DREAMACT_API_KEY"]}, "primaryEnv": "DREAMACT_API_KEY"}}
---
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
Upload your local files to OSS first (get upload policy, then upload).
### 3. DreamAct API
```
POST https://api.newportai.com/api/async/wan/dreamact/2.1
```
### 4. Polling for Result
```
Confidence
79% confidence
Finding
The skill explicitly instructs users to upload local files to external object storage and then submit them to a third-party API. Because the skill processes face images and driving videos, this creates real data-transfer and privacy risk: potentially sensitive biometric/media content leaves the local environment and is stored or processed by external infrastructure.

External Transmission

Medium
Category
Data Exfiltration
Content
```
### 4. Polling for Result
```
POST https://api.newportai.com/api/getAsyncResult
```

## Implementation Notes
Confidence
75% confidence
Finding
This endpoint is used to poll asynchronous job results from an external service, confirming ongoing transmission of job metadata and retrieval of generated outputs from a third party. In this skill's context, the outputs are derived from user-supplied face images and videos, so the external processing of biometric-like media makes the transmission security-relevant rather than a harmless reference.

Static analysis

No suspicious patterns detected.