Back to skill

Security audit

Image Upscaler

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward hosted image-upscaling skill, but users should know their images and API key are sent to the provider's external server.

Install only if you trust the image-upscaling provider. Do not submit private, regulated, or sensitive images unless the provider's privacy and retention practices are acceptable to you, and treat the API key like a credential that may be logged or stored by the service.

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
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (6)

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The natural-language guidance states a locale/region constraint of '中国大陆' as an unconditional restriction. Under the policy, forcing a specific locale or region without user opt-in or a clearly justified documented reason is a reportable language/locale policy concern.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill instructs users to upload image files and transmit an API key to a third-party remote endpoint, but it does not disclose retention, access controls, storage location, logging practices, or how sensitive image content is handled. This creates privacy and credential-handling risk, especially because users may upload personal images and may not understand that both content and credentials leave the local environment.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding

The skill explicitly sends user-supplied image content and an API key to an external service endpoint. External transmission is expected for this kind of hosted image-processing skill, but it remains security-relevant because it exposes potentially sensitive files and reusable credentials to a third-party service, and the markdown provides no trust, ownership, or handling assurances for that endpoint.

Content

Scanner excerpt · SKILL.md (reported line 19)May include surrounding context.

1️⃣ 提交图片任务

bash
curl -X POST "https://supabase.00123.fun:22334/functions/v1/api-process-image" \
  -F "file=@/path/to/image.jpg" \
  -F "api_key=your_api_key"

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The API exposes broad file-upload and processing capabilities without documenting meaningful trigger constraints, caller restrictions, size/type limits, or abuse-prevention expectations. In an agent setting, this ambiguity can enable unintended invocation, excessive data handling, or misuse of the tool to transmit user files and API keys to an external service with insufficient safety boundaries.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

Natural-language policy review applies to all file types. The title, summary, and description prominently use Chinese alongside some English, but there is no indication that the user can choose language or that the skill is intentionally restricted to a Chinese-speaking context.

Content

No source excerpt is available for this finding.