ClawHub
Back to skill

Security audit

Project Management

Security checks across malware telemetry and agentic risk

Overview

This documentation-only skill coherently describes using CellCog for project workspaces, document upload, AI summaries, and share links, with sensitive behaviors mostly disclosed.

Install only if you intend to use CellCog as a remote document/project service. Do not upload confidential, regulated, or third-party files unless you have permission, and treat signed URLs like temporary secrets because anyone with the link can access the document until it expires.

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
92% confidence
Finding
The skill promotes generation of signed URLs that bypass normal CellCog authentication and explicitly notes they remain valid even if project access is later revoked, but it does not clearly warn users that this is an external data-sharing action. In an agent setting, this can lead to unintended disclosure of sensitive project files to other agents, tools, or humans through bearer-style links that are difficult to retract once issued.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The upload flow instructs users to send local documents for remote AI processing, including extraction and summarization, but does not clearly warn that file contents leave the local environment and are transmitted to a third-party service. In practice, agents may upload sensitive internal documents under the assumption this is ordinary local processing, creating confidentiality and compliance risks.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.