Back to skill

Security audit

heating-setback-optimizer

Security checks for vulnerabilities and agentic risk

Overview

This local heating calculator has proportionate access, but its cost estimates and suggested temperatures need independent checking.

Use as a rough estimate. Zero savings can conceal modeled extra costs, so verify heat-pump results independently. Set safe temperature limits before applying schedules; consult [Red Cross freeze-protection guidance](https://www.redcross.org/get-help/how-to-prepare-for-emergencies/types-of-emergencies/winter-storm/frozen-pipes.html).

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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 (5)

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

Unrestricted electric-baseboard setback advice omits a freeze-protection warning.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The claim that electric-baseboard setback 'always pays, any depth' turns an energy-efficiency observation into unrestricted operational advice without a freeze-protection constraint. Because this reference informs thermostat scheduling, an agent could recommend excessively low temperatures during freezing weather, risking frozen pipes, flooding, and unsafe indoor conditions; no direct thermostat-control capability is established by this file.

Content

No source excerpt is available for this finding.

Negative net savings are clamped to zero, hiding the increased energy cost of setbacks that backfire.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

After subtracting the heat-pump recovery penalty, max(0.0, saved) discards negative net savings, causing increased energy consumption and cost to appear as zero savings in both individual results and scan rankings. This undermines the skill's explicit purpose of identifying setbacks that backfire and could lead users to adopt unnecessarily expensive heating schedules. The generic strip-heat warning and separately reported penalty partially mitigate the issue but do not disclose the calculated net loss.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.