ClawHub

Long Research

Security checks across malware telemetry and agentic risk

Overview

This research skill is transparent about its purpose, but it needs review because it can run long autonomous browser research using stored browser sessions, login-gated pages, and remote browser infrastructure.

Install only if you are comfortable with long-running browser automation. Use Interactive mode first, remove remote browser mode for sensitive work, use dedicated disposable browser-use profiles, avoid personal logged-in sessions, and require explicit approval before accessing any authenticated site.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Vague Triggers

Medium
Confidence
87% confidence
Finding
The activation examples include broad phrases like "research X" and especially "pull up long research," which could plausibly match ordinary user requests and trigger the skill unexpectedly. Because this skill immediately enters a mandatory workflow, pushes tool usage, file writes, and possible sub-agent delegation, accidental activation can cause unintended external actions and disclosure of sensitive user queries to tools or providers.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill instructs users to create persistent browser profiles, accept cookie consent, and reuse those profiles across sessions without warning that cookies and session artifacts will be stored locally. In a long-running research skill that visits many sites, this can lead to unintended retention of tracking identifiers, login state, or other session data, increasing privacy risk and the chance of cross-site contamination or accidental use of authenticated state.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The guidance explicitly recommends using browser automation for login-gated forums without any guardrails about authorization, user consent, account handling, privacy, or site terms. In a research automation skill, this normalizes accessing restricted content and can lead agents or users to misuse credentials, collect private data, or violate platform policies.

Ssd 4

High
Confidence
96% confidence
Finding
The workflow escalates from ordinary search and fetch steps to instructions like 'crack any login-gated forums,' which frames restricted areas as targets to overcome rather than boundaries to respect. Because this skill is designed for long-running autonomous research with browser tooling, the language materially increases the chance of unauthorized access attempts, privacy violations, and collection of non-public information at scale.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal