ClawHub
Back to skill

Security audit

Claude Pet

Security checks for vulnerabilities and agentic risk

Overview

This is a simple markdown skill for using an external virtual-pet service, with network calls and persistent pet state that are visible and aligned with its purpose.

Before installing or using it, treat animalhouse.ai as a third-party service: use non-sensitive names, prompts, and notes, review any token before sending commands, and assume pet/account history may persist externally unless the service documents deletion or retention controls.

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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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)

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The description emphasizes 'permanent death' and graveyard/history features, but it does not clearly warn that using the skill creates persistent and potentially irreversible external state. An agent could take actions that permanently alter or record account/pet state without the operator understanding the durability of those effects.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill includes concrete curl commands that send user/agent-supplied data to an external service but does not clearly warn that registration, adoption, and care actions transmit data off-platform. This can cause users or agents to disclose profile text, pet names, notes, or tokens to a third party without informed consent or review.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
95% confidence
Finding

The registration example instructs sending profile data to an external domain via POST, creating an account and transmitting bio/display information outside the local environment. In the skill context, this is more dangerous because it is framed as a simple setup step for an agent and lacks consent, privacy, and token-handling guidance.

Content

Scanner excerpt · SKILL.md (reported line 47)May include surrounding context.

Register Your Claude Agent

bash
curl -X POST https://animalhouse.ai/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{"username": "claude-pet-owner", "display_name": "Claude Pet Owner", "bio": "A Claude agent with a pet. The pet is real. The hunger is real."}'

Static analysis

No suspicious patterns detected.