Security audit

Stiff-Sec — OpenClaw Security Hardener

Security checks for vulnerabilities and agentic risk

Overview

This security hardener is purpose-aligned, but it changes persistent OpenClaw configuration while overstating some protections and offering limited user control before writes or restores.

Review the exact changes before running apply, and keep the backup directory private because backups may contain tokens from your OpenClaw config. Do not rely on the advertised checksum, signed lockfile, permission hardening, or tamper detection claims in this version unless you verify and patch the scripts yourself.

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

Findings (2)

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The script overwrites ~/.openclaw/openclaw.json directly after modifying security-relevant settings, without prompting the user or requiring an explicit confirmation step. In an agent-skill context, silent modification of a user's active configuration can cause unexpected policy changes, break workflows, or lock users into settings they did not approve.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The restore path copies the most recent backup over the active configuration with no confirmation, which can silently revert security settings or operational state. In this skill context, automatic replacement of a user's live config is more dangerous because it manipulates persistent agent behavior and may undo intentional hardening or re-enable previously risky settings.

Static analysis

No suspicious patterns detected.