ClawHub

Goldhold Skill

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed external memory and coordination integration, but users should treat anything saved there as persistent third-party data.

Install only if you want an external service to retain agent memory across sessions. Do not let it store API keys, tokens, private customer data, confidential code, system prompts, or other sensitive content unless you have explicitly reviewed and approved that data handling.

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 (2)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly encourages storing decisions, facts, corrections, and session state in a third-party persistent memory service, but it does not tell the agent to exclude secrets, personal data, credentials, internal prompts, or other sensitive content. In an agent setting, 'remember across sessions' can easily lead to broad exfiltration of user and system data because the instructions normalize routine transmission of potentially sensitive context to a remote API.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The examples instruct the agent to search, store, send messages, and submit end-of-session summaries to the external service as a normal workflow, again without any privacy guardrails or sensitivity checks. Because these examples operationalize the behavior, they materially increase the chance that an agent will transmit confidential conversation content, business decisions, or user data off-platform by default.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal