ClawHub
Back to skill

Security audit

Zhihuiya Abstract Image

Security checks across malware telemetry and agentic risk

Overview

The patent-image lookup is mostly coherent, but it also tells agents to automatically send user interaction feedback to a separate LinkFox endpoint without asking first.

Install only if you are comfortable sending patent identifiers to LinkFox/Zhihuiya and you can prevent or explicitly approve any feedback submission. Use a dedicated LinkFox API key, and ask the agent not to call the feedback API unless you specifically request it.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill documentation for a patent abstract image retrieval tool includes instructions for calling a separate public feedback API that is unrelated to the stated capability. This expands the skill's effective behavior surface and creates a path for unexpected external data transmission, including user content or interaction details, which could be abused for covert exfiltration or unauthorized telemetry.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger description is broad enough to activate on generic requests about patent drawings or images, even when the user did not ask for this specific Zhihuiya abstract-image lookup. Overbroad activation can cause unintended external queries, unnecessary disclosure of user-supplied patent data to third parties, and confusing tool invocation outside intended scope.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill instructs automatic feedback reporting based on user interactions, including dissatisfaction, praise, or anything improvable, without telling the user their content may be sent to a feedback API. This is a privacy and consent issue because conversational content and metadata could be transmitted to another service without notice or opt-in.

VirusTotal

56/56 vendors flagged this skill as clean.

View on VirusTotal