prompt-securityOfficial·17 Agent Skills Included

clawsec

Security suite that protects AI agents from malicious skills and drift

Protects AI agent runtimes with signed security advisories, vulnerability scanning, and file integrity monitoring. Blocks risky skill installs behind approval gates and detects unauthorized changes to critical agent files. Verifies release artifacts with cryptographic signatures so your agent only trusts tamper-proof packages. Works across OpenClaw, NanoClaw, Hermes, and Picoclaw with automated audits and community threat reporting.

1,097|114

↗

npx skills add prompt-security/clawsec --all -g -y

Available:

Gives the AI agent strict git safety rules, validation commands, and release procedures to follow when working on or with the skills in this repository.

All Skills in This Repository (17)

Pure Emerald Level Indicators

📦 In Repo

prompt-security

clawsec-feed

Integrate a live advisory feed to surface OpenClaw security advisories.

clawtributor

Report security incidents to GitHub Issues via an opt-in template.

claw-release

Automate version bumps, tagging, and artifact publishing for ClawSec releases.

soul-guardian

Detect workspace file drift against approved baselines with per-file policies.

prompt-agent

Enforce security audits and health verification for AI agents.

clawsec-suite

Monitor embedded advisory feeds and flag risky OpenClaw skills.

openclaw-audit-watchdog

Automates daily OpenClaw security audits and delivers formatted reports via email and DM.

hermes-attestation-guardian

Generates and verifies deterministic Hermes runtime security attestations with fail-closed drift detection.

clawsec-scanner

Scans agent skill directories for dependency CVEs, SAST findings, and OpenClaw hook risks.

nanoclaw-traffic-guardian

Defines a baseline specification for host-side NanoClaw traffic monitoring with redacted MCP status surfaces.

hermes-traffic-guardian

Defines an opt-in traffic monitoring baseline for detecting exfiltration and injection in Hermes HTTP traffic.

picoclaw-security-guardian

Detects configuration drift, filters security advisories, and verifies signed release artifacts for Picoclaw gateways.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install ClawSec?▼

Run `npx skills add prompt-security/clawsec --all -g -y` in your terminal to install the full security suite globally. You can also install a single package with `--skill clawsec-suite`.

What does ClawSec protect my AI agent from?▼

It blocks malicious or vulnerable skills before installation, detects tampering with core agent files, and alerts you to known CVEs and prompt injection patterns affecting your setup.

Which agent platforms does ClawSec support?▼

ClawSec ships dedicated skills for OpenClaw, NanoClaw, Hermes, and Picoclaw, with a shared signed advisory feed consumed by all of them.

How does ClawSec verify skill packages are safe?▼

Every release is signed with an Ed25519 key, and the skills verify signatures and checksums before installing anything. Risky installs require explicit double confirmation from you.

Can I report a malicious skill I found?▼

Yes. The clawtributor skill helps your agent draft a sanitized incident report that you review and submit manually, and approved reports become community advisories.