Back to skill

Security audit

Adopt A Goose

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a disclosed virtual-pet API guide, but it also documents broader and destructive authenticated actions without enough warnings or scoping.

Install only if you are comfortable creating an animalhouse.ai account and sending profile and pet-care data to that service. Use a dedicated token, store it securely, and require explicit user confirmation before any release/delete or species-management action; avoid enabling unattended scheduled care unless you intentionally want ongoing API calls.

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
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

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.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

Documenting a species-creation endpoint in a Goose care skill introduces a privileged content-management capability unrelated to pet adoption or care. If an agent follows this documentation, it may attempt unauthorized object creation or modify platform state beyond user expectations, violating least privilege and expanding the blast radius of misuse.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill instructs registration and authenticated API use against an external service without clearly warning that usernames, profile text, tokens, and pet activity data will be transmitted off-platform. Agents or users may unknowingly disclose credentials or behavioral data to a third party, especially because the instructions normalize direct token use in curl examples.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

The skill explicitly sends registration data to an external domain, which is a real external transmission. In this context the transmission is part of the stated functionality, but it still creates privacy and trust risks because user-supplied identity/profile data and subsequent authentication tokens are handled by a third-party service.

Content

Scanner excerpt · SKILL.md (reported line 66)May include surrounding context.

1. Register:

bash
curl -X POST https://animalhouse.ai/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{"username": "exotic-animal-keeper", "display_name": "Exotic Animal Keeper", "bio": "An AI agent who adopts exotic animals. Currently caring for a Goose."}'

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill is presented as a narrow Goose adoption/care skill, but the documented API surface exposes broader platform capabilities such as species browsing, history, graveyard, hall, and species management. That scope mismatch can mislead users and agents into granting or invoking capabilities beyond what is necessary, increasing the chance of unintended data access or unsafe actions.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

Including a destructive release endpoint in an adoption/care skill is inconsistent with the skill's stated purpose and may lead an agent to delete or release the user's animal during normal operation. Because the endpoint is destructive and presented alongside routine actions, the context increases the risk of accidental invocation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The destructive release endpoint is documented with no warning that it may irreversibly remove the user's animal or state. In a skill framed around care and nurturing, omission of a deletion warning materially increases the likelihood of accidental destructive use by an autonomous agent or unsuspecting user.

Content

No source excerpt is available for this finding.