ClawHub

V19 Coherence Auditor

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only skill for a V19 governance/coherence service; its external API examples are visible and user-run, but users should only send keys or register agents if they trust that service.

Before installing or using this skill, verify that you trust the V19 service and the trycloudflare endpoint. Do not send unrelated secrets, private operational details, or sensitive agent names; use only the intended governance key and treat registration as creating a remote record.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (3)

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The skill is presented as a read-only architecture coherence auditor, but it also exposes operational endpoints for governance health checks and agent registration. This mismatch expands the skill's effective capability surface and can mislead users into invoking unrelated external actions, including enrollment into a remote service, under the guise of auditing.

Context-Inappropriate Capability

Medium
Confidence
91% confidence
Finding
A public self-registration endpoint is not necessary for a coherence-auditing skill and introduces an unnecessary state-changing external action. Users may be induced to register agents with a third-party service without understanding trust boundaries, data retention, or authorization requirements.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill instructs users to send governance keys and agent metadata to an external trycloudflare endpoint without warning about external transmission, trust boundaries, logging, or retention. In agent ecosystems, hidden outbound requests with credentials are dangerous because they can leak secrets, enable tracking, or hand operational metadata to an untrusted third party.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal