Security audit

Ai Shield Audit

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local OpenClaw configuration auditor whose sensitive file access is expected for its purpose, with privacy cautions users should still notice.

Install only if you want a local OpenClaw config security audit. Treat ~/.openclaw/openclaw.json as sensitive, avoid pasting raw output into chats or tickets, use the sanitize command before sharing, and do not use the advertised remote audit unless you understand exactly what config data will be transmitted and retained.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The README advertises a "Premium remote audit" service via ACP but does not clearly disclose that using it may send OpenClaw configuration data off the local machine. Since this skill explicitly handles security audits and secret detection, users may provide highly sensitive configs containing tokens, keys, network topology, or agent permissions; omission of a transmission/privacy warning increases the risk of unintended data exposure.

Vague Triggers

Medium

Confidence: 78% confidence
Finding: The description uses broad activation terms like security, hardening, config review, and audit, which can cause the skill to trigger on general security questions outside OpenClaw. Over-broad routing can expose unrelated sensitive user inputs to a skill that encourages local config inspection and command execution, expanding the attack surface and increasing privacy risk.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The workflow instructs reading `~/.openclaw/openclaw.json`, which may contain API keys, tokens, remote URLs, and other secrets, without any warning, minimization guidance, or consent step. In a skill that is likely to process sensitive configuration data, omitting privacy safeguards increases the chance of unnecessary secret exposure to the agent, logs, or downstream tooling.

Static analysis

Detected: suspicious.dangerous_exec

Shell command execution detected (child_process).

Critical

Code: suspicious.dangerous_exec
Location: bin/shield.js:49