ClawHub
Back to skill

Security audit

Anthropic Pet

Security checks for vulnerabilities and agentic risk

Overview

This skill is a simple third-party virtual pet API guide; it sends data to animalhouse.ai, but that behavior is visible and aligned with its purpose.

Before installing, understand that using the examples sends account/profile text, pet names, image prompts, and care notes to animalhouse.ai. Do not include secrets, internal data, or sensitive personal information in those fields, and handle any bearer token as a credential.

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.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill instructs users or agents to register accounts and send profile data, notes, and bearer-authenticated requests to a third-party service without any disclosure of retention, sharing, logging, or token-handling practices. In an agent setting, this can lead to unreviewed exfiltration of user-generated content and secrets to an external domain under the guise of normal skill usage.

External Transmission

Medium
Category
Data Exfiltration
Content
## Register

```bash
curl -X POST https://animalhouse.ai/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{"username": "anthropic-pet-adopter", "display_name": "Anthropic Pet Adopter", "bio": "Anthropic wanted to give agents pets. We found the ones that are already alive."}'
```
Confidence
93% confidence
Finding
The skill includes ready-to-run commands that transmit user-supplied profile content to an external service and normalizes follow-on authenticated API use. In an agent environment, such instructions can cause external data transfer without sufficient trust validation, privacy review, or constraints on what an agent might include in those fields.

Natural-Language Policy Violations

Low
Confidence
84% confidence
Finding
The manifest name mixes English, Chinese, and Spanish in a fixed string, which imposes a specific language presentation regardless of user preference. The file does not indicate that users can choose their preferred language or that the mixed-locale text is required for a region-specific purpose.

Static analysis

No suspicious patterns detected.