ClawHub
Back to skill

Security audit

Research

Security checks across malware telemetry and agentic risk

Overview

This is a coherent research-note skill, but its setup and automation instructions include enough sensitive persistence, shell-profile changes, remote installer execution, and delayed agent activity that users should review it before installing.

Install only if you are comfortable with research prompts and outputs being saved under ~/.openclaw/workspace/research and, for deep research, sent to Parallel AI. Before enabling the setup, verify the referenced helper scripts from a trusted source, avoid curl-to-shell installation where possible, prefer user-local symlinks over sudo, and use a scoped or revocable API key rather than broadly loading secrets from your shell profile.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (5)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The setup instructs users to create a system-wide symlink under /usr/local/bin using sudo and later to run a remotely fetched install script via curl | sh, but it does not warn about the trust and integrity risks of privileged changes or piping network content directly to a shell. In a setup document, these patterns can normalize unsafe installation behavior and increase the chance of compromise if the upstream source is tampered with or the path target is replaced.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger conditions are very broad and overlap with common exploratory requests, so the skill may activate when a user did not intend to start persistent research workflows. Because this skill creates files and can invoke deeper tooling, unintended activation can lead to unnecessary filesystem changes or external operations without clear user awareness.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill directs the agent to create and update files under the user's home workspace, but the description does not foreground that activating the skill will modify the filesystem. Users may believe they are just asking for conversational research when the skill is actually persisting prompts, notes, PDFs, and other artifacts.

Session Persistence

Medium
Category
Rogue Agent
Content
- Research a topic, idea, or question
- Explore something before committing to building it
- Investigate options, patterns, or approaches
- Create a "research doc" or "investigation"
- Run deep async research on a complex topic

## Research Directory
Confidence
86% confidence
Finding
Create a "research doc" or "investigation" - Run deep async research on a complex topic ## Research Directory Each research topic gets its own folder: ``` ~/.openclaw/workspace/research/<topic-slug>/

YARA rule 'backdoor_persistence': Backdoor persistence with malicious payloads (shell commands, SSH key injection, hidden root users) [malware]

High
Category
YARA Match
Content
chmod 600 ~/.secrets/parallel_ai/.env

# Add to your shell profile (~/.bashrc or ~/.zshrc)
echo 'export $(cat ~/.secrets/parallel_ai/.env | xargs)' >> ~/.bashrc
source ~/.bashrc
```
Confidence
93% confidence
Finding
echo 'export $(cat ~/.secrets/parallel_ai/.env | xargs)' >> ~/.bashrc

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.