ClawHub
Back to skill

Security audit

Seasonal Planting Guide

Security checks across malware telemetry and agentic risk

Overview

This is a local gardening calendar skill that only saves custom planting data and exports calendar files when the user asks it to.

Reasonable to install for garden planning. Be aware that adding custom plants saves notes locally under ~/.openclaw/workspace, and using --export can create or overwrite a markdown file at the path you choose, including many locations inside your home directory.

SkillSpector

By NVIDIA
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)

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The path-safety function claims to restrict writes to the workspace, home, or /tmp, but its actual logic allows writes anywhere under the user's home directory except for a short denylist of sensitive paths. This creates a broader write surface than the security comment suggests, so a caller may incorrectly trust that exports are confined to a safer workspace boundary while the tool can still overwrite many user files in home.

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The error message tells users that writes are limited to the workspace or home directory and not system paths, but the underlying behavior also permits /tmp and broadly allows home-directory writes. Misstated security boundaries are dangerous because operators and downstream agents may make unsafe assumptions about where attacker-controlled output can be written, enabling file overwrite or persistence in unexpected user-controlled locations.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.