ClawHub

ClawGuard

Security checks across malware telemetry and agentic risk

Overview

ClawGuard is a coherent security guardrail installer, but users should configure it carefully because it can inspect, log, and send tool-call context to an LLM.

Install only if you trust the @capsulesecurity/clawguard package and are comfortable with a gateway plugin seeing tool calls. In sensitive environments, consider disabling full tool-call logging and metrics, confirm which LLM provider receives evaluation context, use a local or approved model when needed, and redact any gateway tokens shown during troubleshooting.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explains that tool context is sent to an LLM for security evaluation, but it does not present this as a clear privacy warning to the user before installation/configuration. Because tool-call context can contain prompts, arguments, file paths, secrets, or other sensitive operational data, sending it to an external or third-party model provider can create unintended data disclosure risk.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The configuration table shows `logToolCalls` is enabled by default, but the skill does not clearly warn that full tool call JSON may be written to gateway logs. Tool call payloads often contain sensitive prompts, parameters, credentials, paths, or business data, so default logging can create a durable secondary exposure surface.

Missing User Warnings

Low
Confidence
86% confidence
Finding
Anonymous usage metrics are enabled by default, but the skill does not clearly call this out as a privacy-relevant behavior. Even when labeled anonymous, telemetry can still be sensitive in regulated or high-security environments, especially if users are not informed before installation.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal