ClawHub
Back to skill

Security audit

mcpjungle

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only admin skill for MCPJungle, and its powerful commands are visible and aligned with that purpose.

Install only if you intend to let the agent help administer an MCPJungle registry. Confirm the registry URL, server, tool, or prompt name before running state-changing commands, and avoid passing secrets or sensitive data into tool invocations unless you trust the target server.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

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
92% confidence
Finding
The skill explicitly advertises destructive management actions such as deregistering, disabling, and invoking components, but provides no guardrails about confirming user intent, validating targets, or warning about service disruption. In an agent context, this increases the chance that operationally dangerous commands are executed on production infrastructure without adequate human awareness.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The documented invoke command sends arbitrary JSON input to a live MCPJungle registry/server, but the skill does not warn that provided data may be transmitted to external tools or downstream services. This can lead to unintended disclosure of secrets, sensitive prompts, or user data when an agent passes unreviewed input to registered tools.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
### Disable / enable prompts
```bash
mcpjungle --registry <url> disable prompt "<server-name>__<prompt-name>"
mcpjungle --registry <url> enable prompt "<server-name>__<prompt-name>"

# Entire server prompts
Confidence
85% confidence
Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
### Disable / enable prompts
```bash
mcpjungle --registry <url> disable prompt "<server-name>__<prompt-name>"
mcpjungle --registry <url> enable prompt "<server-name>__<prompt-name>"

# Entire server prompts
Confidence
85% confidence
Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Static analysis

No suspicious patterns detected.