ClawHub
Back to skill

Security audit

Miremo Research Skill

Security checks across malware telemetry and agentic risk

Overview

This skill coherently connects OpenClaw to a user's Miremo knowledge base, with disclosed read/search and note-creation capabilities that match its stated purpose.

Install this only if you want OpenClaw to access your Miremo account through an API key. Keep the key out of version control and synced dotfiles, revoke it if exposed, and be aware that the skill can read your Miremo content and create new memos when you ask it to save information.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Missing User Warnings

Low
Confidence
84% confidence
Finding
The README instructs users to place a live Miremo API key directly into a local OpenClaw config file, but it does not explicitly warn about risks such as accidental sharing, weak file permissions, shell history leakage during editing, or inclusion in backups and dotfile sync tools. This is not inherently malicious, but it does normalize storing long-lived credentials in plaintext and can lead to credential exposure if the local environment is compromised or the config is mishandled.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill exposes a write-capable tool (`create_memo`) and encourages its use for phrases like 'save/record/remember this', but it does not clearly require explicit user confirmation immediately before performing the write. In an agent setting, ambiguous requests or prompt injection in retrieved content could cause unintended persistence of sensitive, incorrect, or adversarial data into the user's knowledge base.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.