ClawHub
Back to skill

Security audit

houseplant-watering-computer

Security checks for vulnerabilities and agentic risk

Overview

This is an offline houseplant watering helper with minor accuracy and safety caveats, but no hidden data access, persistence, or unrelated authority.

Installers should expect an offline Python helper for plant-watering estimates, not a full scheduling/card system. Verify any plant-care advice before applying it broadly, especially the dish-soap suggestion for hydrophobic substrate; safer options include repeated plain-water soaking or a horticultural wetting agent.

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)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The code substantially matches the declared gardening/watering purpose: it computes watering intervals and dose from substrate, pot geometry, plant class, light, temperature/VPD, humidity, and season; it also includes moisture checking, symptom interpretation, and self-watering/cache-pot guidance. However, two declared capabilities are not actually implemented as described. First, the description claims plant size leaf area is part of the computation, but the code only uses fixed plant classes and a substrate-volume scaling exponent, not explicit leaf area or plant-size input. Second, the description says it generates a per-pot care card with moisture-check dates, but the code only emits JSON with check intervals and expected interval ranges, not a formatted care card or dated schedule. There are no suspicious undeclared permissions or unrelated behaviors; the mismatch is due to overclaiming some features.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The symptoms guidance recommends adding a drop of dish soap to a soak basin for hydrophobic substrate without a clear warning that soap formulations vary and may harm sensitive plants, roots, or soil microbiota. Although framed as practical plant-care advice, it is still a real safety issue because users may apply unsuitable products directly to plant media based on the tool's recommendation.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.