Enterprise Agent Security Framework — Credential scanning, fake agent detection, egress controls, trust scoring, and compliance automation for AI agent deployments. Version 1.1 includes security hardening, improved YARA scanning, and enhanced neuro-symbolic verification.
Version: 1.1.0
Category: Security / DevOps / AI Agents
Platform: OpenClaw
License: Apache 2.0
ASF (Agent Security Framework) is an enterprise-grade security orchestration framework for AI agent deployments:
# Install via curl
curl -s https://raw.githubusercontent.com/jeffsutherland/asf-distro/main/install.sh | bash
# Or clone and run manually
git clone https://github.com/jeffsutherland/asf-distro.git
cd asf-distro
./install.sh
Version 1.1 focuses on security hardening and operational resilience:
# Verify ASF installation
asf verify
# Scan a skill for security issues
asf scan --skill ./my-skill
# Check agent trust score
asf trust-score --agent <agent-id>
# Run compliance check
asf compliance --framework SOC2
| Feature | Description |
|---|---|
| Docker-Native | Official image: docker.io/jeffsutherland/asf-framework:v3.2 |
| Neuro-Symbolic v2 | Hybrid reasoning + symbolic logic verification |
| Tiered Egress L0–L4 | Granular outbound traffic control |
| Compliance Matrix | SOC 2 (92%), HIPAA (95%), FedRAMP (88%), PCI-DSS (85%) |
| Grok Integration | /asf-scan command on Grok marketplace |
Configuration files in configs/:
docker-security.json — Docker security profile settingsegress-whitelist.json — Egress allowlist rulescredential-vault.json — Credential management configurationFull documentation at: https://sandbox.jvsmanagement.com/agentsecurityframework/
ASF — Secure your AI agents, starting at the foundation.