ClawHub
Back to skill

Security audit

fulcra-get-started

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed onboarding guide for connecting to Fulcra, but users should be aware it can connect agents to sensitive personal data and remote storage.

Before installing, understand that this skill helps an agent authenticate to Fulcra, use a remote data platform, and store or retrieve files there. Only connect data sources you intend to share with Fulcra, avoid uploading secrets or regulated personal data unless you explicitly want that, and remember the CLI stores refreshable credentials locally.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The file-store section instructs agents to upload and download files to a remote Fulcra service but does not explicitly warn that this transfers potentially sensitive user data off-host. In an agent skill, omission of a consent/sensitivity check can lead to unintended exfiltration of personal notes, artifacts, or other regulated content because the agent is encouraged to operationalize these commands directly.

Natural-Language Policy Violations

Low
Confidence
78% confidence
Finding
The guidance says the user's local time zone can be inferred from system time, which permits use of ambient environment data without user opt-in. While low severity, this can create privacy leakage and incorrect assumptions in shared, server-side, or remote execution environments where system timezone may reveal host context rather than the user's preference.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.