ClawHub
Back to skill

Security audit

Luma Ai

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward Luma AI video-generation helper, with a visible third-party API example that users should treat with normal privacy and API-key care.

Before using the API example, understand that prompts, images or references, authentication headers, and generated asset metadata may be sent to Luma's external service. Use environment variables or a secret manager for API keys, avoid sensitive or regulated content unless approved, and review Luma's terms and privacy practices.

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 (6)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill includes sample code that sends user prompts and an API bearer token to a third-party service but does not warn users that their content will leave the local environment. In a skill context, this can cause unintentional disclosure of sensitive prompts, generated asset metadata, or mishandling of credentials by users who copy-paste the example without understanding the privacy implications.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

This duplicate finding points to the same POST-based external transmission behavior: prompts and authorization data are sent to Luma's API. The main risk is not malicious code execution but inadvertent data disclosure and poor credential hygiene if copied directly from the documentation.

Content

Scanner excerpt · SKILL.md (reported line 66)May include surrounding context.

python
import requests

response = requests.post(
    "https://api.lumalabs.ai/dream-machine/v1/generations",
    headers={"Authorization": "Bearer luma-xxx"},
    json={

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

This duplicate finding points to the same POST-based external transmission behavior: prompts and authorization data are sent to Luma's API. The main risk is not malicious code execution but inadvertent data disclosure and poor credential hygiene if copied directly from the documentation.

Content

Scanner excerpt · SKILL.md (reported line 66)May include surrounding context.

python
import requests

response = requests.post(
    "https://api.lumalabs.ai/dream-machine/v1/generations",
    headers={"Authorization": "Bearer luma-xxx"},
    json={

External Transmission

Medium
Category
Data Exfiltration
Confidence
92% confidence
Finding

The hardcoded external endpoint confirms that the skill directs users to send content to a third-party domain. In context, this is consistent with a Luma integration skill, so it is less suspicious than covert exfiltration, but it still presents privacy and compliance risk when no consent or data-sharing notice is provided.

Content

Scanner excerpt · SKILL.md (reported line 67)May include surrounding context.

md
import requests

response = requests.post(
    "https://api.lumalabs.ai/dream-machine/v1/generations",
    headers={"Authorization": "Bearer luma-xxx"},
    json={
        "prompt": "a cat sitting on a windowsill, warm sunlight",

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

The polling GET request sends the bearer token to the same third-party API and retrieves generation results, potentially including URLs to generated assets. While normal for the feature, it extends the external data flow and credential exposure surface, especially if users reuse insecure code patterns or do not understand that asset metadata and links are obtained from an external service.

Content

Scanner excerpt · SKILL.md (reported line 79)May include surrounding context.

md
# 轮询获取结果
result = requests.get(
    f"https://api.lumalabs.ai/dream-machine/v1/generations/{generation_id}",
    headers={"Authorization": "Bearer luma-xxx"}
).json()
# result["assets"]["video"] 为视频 URL

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
85% confidence
Finding

The statement that English prompts usually work better introduces a language preference in the skill's guidance without explicitly presenting it as an optional recommendation or offering a user language choice. The policy for SQP-3 calls out language or locale constraints that are imposed or steered without opt-in.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.