ClawHub

Tarkov API + Wiki Hardcore Assistant

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Tarkov game-data lookup skill with expected network calls and user-directed local file inputs, but users should avoid raw mode and unsafe endpoints unless they intentionally review them.

Use the preset commands for normal gameplay lookups. Only provide minimal stash JSON/CSV files you intend the tool to read, and avoid raw mode, query files, custom endpoints, or --allow-unsafe-endpoint unless you have reviewed the query and trust the destination.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding
The skill explicitly relies on network access, local file input, and likely environment access via its Python script, but the skill file does not declare permissions. That creates a governance gap: users and the platform cannot easily review or constrain what the skill is allowed to access, increasing the chance of unintended data exposure or overbroad execution.

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The raw subcommand allows arbitrary GraphQL queries and variables to be sent to the remote API, which expands the tool far beyond the curated, bounded operations described in the skill metadata. In an agent setting, this creates a capability-exposure issue: prompt-influenced users or downstream chains can invoke unintended data access patterns or expensive queries, bypassing the safer preset commands.

Intent-Code Divergence

Medium
Confidence
89% confidence
Finding
The tool presents itself as a secure helper, but simultaneously exposes raw query execution and a flag to disable endpoint safety checks. That mismatch increases the chance that orchestrators or users will trust the tool more than they should, enabling abuse paths that the description implies are prevented.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal