ClawHub
Back to skill

Security audit

Media.io AI Image & Video Generation API

Security checks across malware telemetry and agentic risk

Overview

This skill is a Media.io API wrapper whose main risk is expected: it sends your prompts, media URLs, and API key to Media.io to generate images or videos.

Install only if you intend to use Media.io's API. Use a dedicated or test API key, monitor credit usage, and avoid submitting sensitive prompts, private media URLs, or proprietary videos unless you are comfortable sharing them with Media.io.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Lp3

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding
The skill documentation and examples indicate access to environment variables (`API_KEY`), local files (`scripts/c_api_doc_detail.json`), and outbound network calls to `https://openapi.media.io`, but these capabilities are not explicitly declared as permissions. This is a real security issue because undeclared capabilities reduce transparency for reviewers and users, making it easier to overlook secret handling and data exfiltration risks, especially in a community-maintained skill that transmits credentials to an external service.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The reference documents a 'Motion Control Kling 2.6' endpoint that accepts both a video and an image, which is effectively a video-to-video style capability not disclosed in the manifest metadata. This creates a scope/transparency mismatch: users and governance systems may believe the skill only handles image/video generation as declared, while it can also transmit user-supplied video assets to a third-party service.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The document broadly describes invoking many third-party Media.io endpoints with API keys and user prompts/media, but does not warn that supplied text, image URLs, video URLs, task IDs, and credentials are sent to an external provider. In a skill context, this lack of disclosure can cause unintended data sharing and privacy/compliance violations even if the network use itself is expected.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The documentation encourages sending user-supplied prompts and remote image/video URLs to third-party Media.io endpoints using an API key, but it provides no privacy, consent, or data-handling warning. In a skill context, this can cause users or integrators to transmit sensitive media, prompts, or internal URLs to an external service without understanding the exposure, creating privacy and SSRF-adjacent risk if arbitrary URLs are accepted upstream.

External Transmission

Medium
Category
Data Exfiltration
Content
{
      "title": "create video by text or image",
      "language": "cURL",
      "code_example": "curl --request POST \\\n  --url https://openapi.media.io/generation/kling/i2v-motion-control-kling-v2-6 \\\n  --header 'Content-Type: application/json' \\\n  --header 'X-API-KEY: <api-key>' \\\n  --data '\n{\n  \"data\": {\n    \"image\": \"<string>\",\n    \"prompt\": \"<string>\",\n    \"video\": \"<string>\"\n  }\n}'\n"
    }
  ],
  "describe": "",
Confidence
90% confidence
Finding
curl --request POST \\\n --url https://openapi.media.io/generation/kling/i2v-motion-control-kling-v2-6 \\\n --header 'Content-Type: application/json' \\\n --header 'X-API-KEY: <api-key>' \\\n --da

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal