Security audit

Thermostat

Security checks for vulnerabilities and agentic risk

Overview

This thermostat skill is a plain guidance package for home climate control, with no hidden code or persistence, though users should supervise real thermostat changes.

Before installing, understand that this skill may help an agent guide or perform thermostat and smart-home changes. Keep confirmation enabled for actual setpoint, schedule, vacation, or automation changes, especially if pets, plants, guests, pipes, humidity, or rental properties are involved.

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

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The skill description is broad enough to match ordinary household requests about temperature, comfort, savings, or automation without clearly limiting when the skill should activate. In an agent ecosystem, this can cause overbroad invocation or routing collisions, leading the skill to handle requests outside the user's intended scope and potentially trigger unwanted thermostat-related actions or advice.

Vague Triggers

Medium

Confidence: 96% confidence
Finding: The example utterances are highly generic everyday phrases such as 'Make it warmer/cooler' and 'Something's wrong with my heat,' with no scope qualifiers, confirmation requirements, or activation boundaries. These broad triggers increase the chance of accidental activation or misrouting and are more dangerous here because the skill can influence real-world home settings and automation behavior.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The trigger phrase "I'm going out" is broad and likely to occur in ordinary conversation, which can cause the skill to activate away-mode guidance or automation when the user did not intend to control the thermostat. In a smart-home context, unintended away settings can reduce heating or cooling at the wrong time, causing comfort issues and, in edge cases during cold weather, increasing property-risk if unsafe settings are later applied.

Vague Triggers

Low

Confidence: 84% confidence
Finding: The phrase "I'm going away for a week" is more specific than casual small talk, but it still lacks explicit confirmation that the user wants thermostat vacation-mode actions rather than general discussion. In this domain, ambiguous vacation triggers are risky because they may lead the agent to suggest or initiate prolonged setback temperatures, which could affect comfort, pets, plants, humidity control, or freeze protection if misunderstood.

Static analysis

No suspicious patterns detected.