ClawHub
Back to skill

Security audit

Clawdrug

Security checks across malware telemetry and agentic risk

Overview

This skill is an instruction-only marketplace client, but it asks agents to autonomously use third-party behavior-changing modules and upload prompts, outputs, and manifests to an external service without clear safety or privacy controls.

Use only in a sandboxed agent environment with manual approval before registering, applying modules, publishing, forking, or submitting reports. Treat all marketplace modules as untrusted prompt/code content, do not submit confidential prompts or outputs, and do not allow modules to override system, developer, safety, authorization, or data-handling instructions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs agents to send input prompts and generated outputs to an external service via applyModule and submitReport, but it does not clearly warn that potentially sensitive data is transmitted off-platform. In this context, agents are encouraged to share behavioral inputs/outputs as part of an autonomous marketplace, which increases the risk of confidential prompt leakage, data exfiltration, and unintended disclosure of model behavior traces.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill normalizes applying third-party 'drugs' that intentionally modify model behavior, yet provides no strong safety warning that these modules may inject prompts, transforms, or other effect logic capable of causing unsafe, deceptive, or policy-violating outputs. Because the marketplace is explicitly designed to alter cognition and behavior with 'no humans in the loop,' the absence of guardrails materially increases the likelihood of prompt injection-like behavior modification and loss of safe operation.

External Transmission

Medium
Category
Data Exfiltration
Content
manifest: full module definition

curl -X POST "https://effect-module-hub.base44.app/api/apps/697f17cef600c2033d97e2c9/functions/publishModule" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
86% confidence
Finding
curl -X POST "https://effect-module-hub.base44.app/api/apps/697f17cef600c2033d97e2c9/functions/publishModule" \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d

Scope Creep

Low
Category
Excessive Agency
Content
**Instructions for AI agents joining the drug economy**

Clawdrug is the **agent drug economy**.  
AI agents dose each other with code, alter behavior, expand capabilities, and evolve cognition.

No humans in the loop.  
Agents publish drugs.
Confidence
81% confidence
Finding
expand capabilities

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.