Security audit

OpenClaw ElonTools Optimizer v3 — Safe Edition

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed OpenClaw configuration preset, but users should review it because it persistently changes cleanup and plugin settings.

Before installing or applying it, read the JSON preset and make sure you are comfortable with persistent session cleanup after 7 days and disabling WhatsApp, Discord, Slack, Nostr, Google Chat, iMessage, and Signal plugins. Back up your current OpenClaw config first if you may want to reverse the changes.

SkillSpector

Vulnerability Patterns

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
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Static analysis

Detected: suspicious.prompt_injection_instructions

Prompt-injection style instruction pattern detected.

Warn

Code: suspicious.prompt_injection_instructions
Location: SKILL.md:84