Security audit

generator-sizer

Security checks for vulnerabilities and agentic risk

Overview

The skill can recommend an undersized generator because its sizing and fit checks ignore the starting power it calculates.

Treat sizing and positive fit results as unverified running-load estimates. Confirm startup compatibility against actual appliance and generator specifications before relying on recommendations, especially for medical equipment, heating, or pumps. Starting and running requirements are distinct, as illustrated in [Honda's sizing guide](https://powerequipment.honda.com/generators/generator-how-much-power).

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

MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
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 (2)

Description-Behavior Mismatch

High

Category: MCP Tool Poisoning
Confidence: 99% confidence
Finding: The skill promises surge-based generator sizing, but the reported implementation uses running load plus headroom for selection and fit verdicts without enforcing starting-surge requirements. This is a safety-relevant mismatch: users preparing backup power for medical equipment, heating, or sump pumps may trust an undersized generator recommendation and lose essential power when a motor starts.
Content: No source excerpt is available for this finding.

Generator recommendations and fit verdicts ignore starting surge despite the manifest promising surge-based sizing.

High

Category: Excessive Agency
Confidence: 99% confidence
Finding: The script computes worst_surge but bases generator selection, fit verdicts, and load shedding only on running watts with 15% headroom. For example, the catalog's sump pump requires 2,700 W to start, yet its 1,035 W continuous requirement selects the 2,000 W class; startup can therefore trip or stall the generator and interrupt essential outage loads.
Content: No source excerpt is available for this finding.