Back to skill

Security audit

image-upscaler

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious because scanners report privacy and sharing risks, while part of the analysis was incomplete.

Before installing, consider whether you are comfortable with images and dataset fields being sent to external services or with sharing permissions being broadened. The scanner reports that coverage is partial: one component was only partly inspected, and some instruction checks could not be fully evaluated. The suspicious verdict signals review-worthy risks, not a finding that the author acted maliciously.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

The workflow sends user images and associated dataset fields to external services without an explicit privacy warning.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The workflow defaults to external image processing and permits forwarding selected dataset fields without explicitly explaining the disclosure or checking authorization for sensitive inputs. Although the skill identifies the providers and offers a local privacy option, its instruction to choose without asking can result in personal photos, confidential images, or associated metadata being sent outside the user's intended trust boundary.

Content

No source excerpt is available for this finding.

The recovery guidance recommends broadening Google Drive sharing without warning about the resulting exposure.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The recovery advice identifies the lack of 'anyone with the link' sharing as the failure cause and suggests fixing sharing without warning that this removes recipient-specific access restrictions. Following that advice for private images could make them accessible to unintended link holders; downloading and re-uploading is offered as an alternative but does not explain the risk of broadening access.

Content

No source excerpt is available for this finding.