ClawHub

Ollama Memory Embeddings

Security checks across malware telemetry and agentic risk

Overview

This skill transparently configures local OpenClaw memory embeddings to use Ollama, with optional watchdog persistence that is documented and user-controlled.

Install this only if you want OpenClaw memory search to use Ollama embeddings. Review the target config path and model choice, enable the watchdog only if you want ongoing background drift-healing, and avoid passing real secrets as --api-key-value. Treat --dry-run cautiously because the installer may still interact with Ollama before printing the plan.

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

Low
Confidence
93% confidence
Finding
The README prominently provides install, enforce, watchdog, and uninstall commands that change OpenClaw configuration and can install persistent launchd automation, but it does not clearly foreground those side effects at the point of execution. In a skill intended for local system integration, this is not inherently malicious, but it is a real safety/usability issue because users may run commands without realizing they are enabling ongoing background remediation or modifying gateway behavior.

Missing User Warnings

Medium
Confidence
98% confidence
Finding
The script logs `apiKey=${APIKEY_NOW}` directly, which can expose credentials in terminal scrollback, CI logs, shell history captures, or centralized logging systems. Even though the default value is `ollama`, the script accepts arbitrary `--api-key-value`, so a real secret could be disclosed if operators supply one.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal