UseAI-proOfficial·13 Agent Skills Included

openclaw-skills-security

Audit skills, scan credentials, and sandbox untrusted code

Audits OpenClaw skills before installation for typosquatting, dangerous permissions, prompt injection, supply chain risks, and data exfiltration. Scans workspaces for leaked API keys and secrets, hardens configurations, and generates Docker sandbox profiles for untrusted code. Includes an incident response playbook for containment, credential rotation, and recovery after a malicious skill is detected.

70|10

↗

npx skills add UseAI-pro/openclaw-skills-security --all -g -y

All Skills in This Repository (13)

Pure Emerald Level Indicators

📦 In Repo

UseAI-pro

permission-auditor

Analyze OpenClaw skill permissions and recommend minimal access sets.

prompt-guard

Detect and neutralize prompt injection attacks in text sources.

setup-auditor

Audit OpenClaw environments for credential leaks, insecure configurations, and missing sandbox isolation.

dependency-auditor

Audit npm, pip, and Go dependencies for vulnerabilities and typosquatting.

network-watcher

Audit OpenClaw skill network requests for exfiltration and unauthorized API calls.

skill-guard

Monitor active OpenClaw skills for runtime security violations.

skill-auditor

Audit OpenClaw skills for security vulnerabilities using a six-step review protocol.

config-hardener

Audit OpenClaw configurations and generate hardened security settings.

sandbox-guard

Generate Docker sandbox configurations with filesystem, network, and process restrictions.

credential-scanner

Scan project workspaces for exposed credentials, API keys, and secrets.

output-sanitizer

Detect and redact credentials, PII, and internal paths from AI agent output.

incident-responder

Guide OpenClaw security breach response through containment, investigation, credential rotation, and recovery.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install openclaw-skills-security?▼

Run `npx skills add UseAI-pro/openclaw-skills-security --all -g -y` in your terminal to install all security skills globally.

How to check if an OpenClaw skill is safe?▼

Load the skill-auditor skill and paste in the target SKILL.md. It runs a six-step review covering permissions, dependencies, prompt injection, and exfiltration, then returns a SAFE, SUSPICIOUS, DANGEROUS, or BLOCK verdict.

How to find leaked API keys in my project?▼

Use the credential-scanner skill to scan your workspace for exposed secrets like AWS keys, OpenAI tokens, and database URLs. It reports findings with masked values and specific cleanup steps.

Does this work with Claude Code and Codex CLI?▼

Yes. All skills follow the standard SKILL.md format and load into Claude Code, Codex CLI, OpenClaw, or any compatible host. You can also paste them directly into any chat.

What should I do after installing a malicious skill?▼

Load the incident-responder skill for a step-by-step playbook covering containment, evidence preservation, credential rotation, and environment recovery.