How does file integrity protection work?

Skills detect drift in critical files and can auto-restore them using stored checksums and verification routines.

Is runtime traffic monitoring fully implemented?

Baseline specifications exist for several platforms, but actual proxy code must be added by platform builders.

Where can I access live security advisories?

Automated NVD CVE polling is built into the suite and accessible via the ClawSec site or installed skills.

prompt-security/clawsec

Security skill suite for protecting AI agents from injections, drift, and threats.

SkillsAgent Skills 1.0kOpen source

View on GitHub

Updated 2026-06-16

prompt-security/clawsec GitHub repository

What is prompt-security/clawsec?

ClawSec functions as a complete security skill suite that adds protection layers to AI agent platforms. It focuses on detecting prompt injections, configuration drift, and supply chain issues through specialized skills that run verification and monitoring tasks.

The tools operate by installing via a central suite mechanism, then applying file checksums, advisory polling from NVD sources, and baseline traffic specs where supported. Platform-specific modules handle tasks like signed feed checks for Hermes or container security for NanoClaw.

It targets developers and operators running OpenClaw, NanoClaw, Hermes, or Picoclaw agents who need built-in safeguards without custom security coding.

What you can build with prompt-security/clawsec

Prevent prompt injection attacks

Deploy scanner and audit skills to detect malicious instructions and maintain agent integrity during operation.

Monitor configuration drift

Use file integrity tools to automatically detect and restore changes to key files like SOUL.md across supported platforms.

Verify supply chain artifacts

Run release verification and checksum skills before installing or updating agent components to block tampering.

Install prompt-security/clawsec

Quick start

# Install the ClawSec security suite
npx clawhub@latest install clawsec-suite

1Clone the ClawSec GitHub repository to your local environment.
2Select and run the suite installer matching your agent platform such as OpenClaw or Hermes.
3Enable the desired skills through the provided installer or integration steps.
4Configure live advisory polling and baseline monitoring options as needed.
5Execute initial security audits and integrity checks to validate the setup.