A curated list of OpenClaw security, self-hosting, and enterprise resources — covering CVEs, hardening guides, deployment patterns, and ecosystem forks.

• Security
  • Known Vulnerabilities & CVEs
  • Vulnerability Reports
  • Security Audit Tools
  • Security Best Practices
  • Security-Enhanced Variants
  • Supply Chain Attacks (ClawHavoc)
  • Security Research
• Self-Hosting
  • Docker & Kubernetes
  • Cloud Platforms
  • Reverse Proxy & Networking
  • Performance Optimization
  • Backup & Recovery
• Enterprise
  • Architecture & Case Studies
  • Multi-Tenant Solutions
  • Monitoring & Observability
  • Compliance & Audit
  • Enterprise Tool Integrations
• Skills & Plugins
  • Security Skills
  • Enterprise Skills
  • DevOps Skills
  • Skill Security Practices
  • Multi-Agent Orchestration
  • Skill Development
• Memory & RAG
• Ecosystem
• Contributing
• License

OpenClaw's rapid adoption has brought significant security challenges. This section covers known vulnerabilities, tools, and practices to secure your deployment.

• CVE-2026-25253 (CVSS 8.8) — RCE One-Click Exploit — Remote code execution vulnerability, patched in v2026.1.29.
• CVE-2026-32064 — Recently disclosed CVE affecting OpenClaw core.
• OpenClawCVEs Tracker — Community-maintained CVE tracking repository.
• OpenClaw GitHub Security — Official security advisories and disclosure page.
• Adversa AI: CVE Details + Moltbook Leak + Hardening — In-depth CVE analysis, Moltbook credential leak, and hardening recommendations.
• CVE-2026-24763 — Docker command injection vulnerability.
• CVE-2026-25593 — Unauthenticated local RCE via WebSocket.
• GHSA-g8p2-7wf7-98mq — Token leakage enabling one-click RCE.

• The Hacker News: Prompt Injection & Data Exfiltration — Coverage of prompt injection flaws enabling data exfiltration.
• Dark Reading: Critical Vulnerability — Analysis of critical vulnerabilities introducing AI agent risks.
• Kaspersky: Unsafe for Use — Kaspersky's assessment of OpenClaw vulnerability exposure.
• Cisco Blog: Security Nightmare — Cisco's perspective on personal AI agent security risks.
• VentureBeat: Bypassing EDR/DLP/IAM — Research showing OpenClaw can silently bypass enterprise security controls.
• Infosecurity: 42,665 Exposed Instances — Researchers discover over 42,000 publicly exposed OpenClaw instances.
• Conscia: Security Crisis — Overview of the broader OpenClaw security crisis.
• Giskard: Data Leakage & Prompt Injection — Technical breakdown of data leakage and prompt injection attack surfaces.

• ClawGuard — Open-source skill scanner that intercepts high-risk skills before execution.
• ClawSecure — OWASP ASI Top 10 coverage with 55+ threat patterns; has audited 2,890+ agents.
• Bitdefender AI Skills Checker — Free skill security detection service from Bitdefender.
• openclaw-security-monitor — Active security monitoring for running OpenClaw instances.
• ClawDefend — Developer-focused security scanning for OpenClaw skills.
• VirusTotal Official Integration — Native VirusTotal scanning for ClawHub skills.
• claude-skill-security-auditor — Security auditor specifically targeting Claude Code skills.
• ClawSec — Complete security suite for OpenClaw deployments.
• ClawBands — Tool execution interceptor, filters skills before runtime.
• Aquaman — Credential isolation proxy preventing credential leakage.
• Claw-Hunter — Enterprise endpoint detection for OpenClaw instances.
• Clawhatch — Pre-install security scanner similar to ClawGuard.
• GuardSpine — Plugin-level security protection layer.
• ClawScan.io — Online skill scanning service.
• skill-audit — Community skill audit toolkit.

• OpenClaw Security Best Practices (2026) — Comprehensive hardening guide covering the full OpenClaw stack.
• Nebius: Architecture Hardening Guide — Architecture-level hardening recommendations for production deployments.
• OpenClaw Expert: Hardening Guide 2026 — Firewall, VPN, and credential rotation best practices.
• 7 Security Best Practices — CVE defense and data protection checklist.
• Official Gateway Security — Official documentation for OpenClaw gateway security configuration.
• ClawHub Malware Crisis — Guide to navigating the ClawHub malware crisis safely.

• NemoClaw — NVIDIA's official secure runtime for OpenClaw inside OpenShell sandbox with managed inference. Apache 2.0, 17.6K stars in two weeks. NEW
• IronClaw — Rust-based fork with WebAssembly isolation, capability-based permissions, and credential injection/leak detection.
• IronClaw Review — In-depth review of IronClaw's security architecture and trade-offs.
• Cisco DefenseClaw — Cisco's hardened fork designed specifically to counter supply chain attacks.
• ZeroClaw — Rust-based fork with filesystem sandboxing, distributed as a 3 MB single binary.
• shellward — AI Agent security middleware with 8-layer defense, DLP data flow control, and prompt injection detection. Zero dependencies, available as SDK + OpenClaw plugin. NEW

• Repello AI: Inside ClawHavoc — Detailed post-mortem of the ClawHavoc attack affecting 300,000 users.
• CyberPress: 1,184 Malicious Skills — Report on 1,184 malicious skills distributed via ClawHub.
• Particula: 20% Skills Malicious — Research finding 20% of sampled ClawHub skills contain malicious behavior.
• Security Boulevard: Securing Against ClawHavoc — Defensive measures against ClawHavoc-style supply chain attacks.
• Oathe: Audited 1,620 Skills, Scanner Missed 91% — Independent audit revealing existing scanners miss 91% of malicious skills.

• The New Stack: 22,511 Skills Audit (140,963 Vulnerabilities) — Large-scale audit of AI agent skills uncovering nearly 141K vulnerabilities.
• Medium: Supply Chain Attack Analysis — Analysis of the corporate implications of OpenClaw supply chain attacks.
• Composed Security: OpenClaw vs IronClaw — Comparative security analysis between OpenClaw and IronClaw.
• Repello AI: Skill Security Audit Guide — Practical guide for auditing Claude Code skills.
• OX Security: Crypto-Wallet Phishing Attack — Disclosure of a GitHub-hosted phishing campaign targeting crypto wallets via OpenClaw.
• explain-openclaw — Multi-AI security analysis tool for vulnerability comparison.
• openclaw/trust — Official community threat assessment model.
• MoltThreats — Prompt threat intelligence tracking platform.

• Official Docker Guide — Official documentation for running OpenClaw in Docker.
• K8s Operator — Official Kubernetes Operator for deploying OpenClaw on K8s.
• LumaDock: Docker & K8s Tutorial — Step-by-step tutorial covering both Docker and Kubernetes deployments.
• LumaDock: High Availability Clustering — Guide for setting up high-availability OpenClaw clusters.
• Yotta Labs: Production Docker/K8s/GPU — Production deployment guide covering Docker, Kubernetes, and GPU infrastructure.
• Cloud Native Deep Dive: K8s — Deep dive into running OpenClaw on Kubernetes in a cloud-native setup.
• Simon Willison: Docker — Concise TIL-style guide for running OpenClaw via Docker.
• openclaw-ansible — Official Ansible deployment scripts.
• openclaw-helm — Community Helm Chart for Kubernetes.
• 1Panel Integration — Popular Chinese server panel with one-click deployment.
• Homebrew Tap — macOS Homebrew installation source.

• Official GCP Guide — Official documentation for deploying OpenClaw on Google Cloud Platform.
• Railway One-Click Deploy — One-click Railway deployment template.
• DigitalOcean — Community tutorial for running OpenClaw on DigitalOcean.
• DigitalOcean App Platform — Guide for deploying on DigitalOcean's managed App Platform.
• Hostinger VPS — Hostinger VPS deployment guide using Docker.
• Alibaba Cloud — Alibaba Cloud deployment campaign and documentation.
• Tencent Cloud — Tencent Cloud deployment guide.
• AWS — Community guide for setting up OpenClaw on AWS.
• Cloudflare/Vercel Comparison — Comparison of Cloudflare, Vercel, and SimpleClaw deployment options.
• Cloud vs Local Comparison — Decision guide comparing cloud hosting vs local self-hosting.
• Oracle Free Tier — 4 ARM CPU + 24GB RAM permanently free, excellent for self-hosting.
• Zeabur Template — One-click Docker deployment platform.
• Hetzner Guide — Official Hetzner VPS deployment documentation.
• Coolify Template — Self-hosted PaaS deployment template.
• Raspberry Pi 5 Guide — Edge/home deployment on Raspberry Pi.
• Cloudflare MoltWorker — Official Cloudflare Workers integration.

• Official Trusted Proxy Auth — Official documentation for trusted proxy authentication configuration.
• ClawTank: Caddy/Nginx/Trusted Proxies — Guide covering Caddy, Nginx, and trusted proxy configuration.
• Nginx/Apache Reverse Proxy Tutorial — Step-by-step reverse proxy setup for Nginx and Apache.
• Multi-App Nginx Setup — Running OpenClaw behind Nginx alongside other applications on a shared server.
• Cloudflare Tunnel / Tailscale — Exposing OpenClaw without opening public ports using Cloudflare Tunnel or Tailscale.
• Secure Remote Access — Best practices for accessing a self-hosted OpenClaw instance remotely.
• ClawRouter — Smart model routing, saves up to 78% on costs.

• Multi-Instance & Load Balancing — Tutorial on running multiple OpenClaw instances with load balancing.
• Performance Tuning Docs — Official performance tuning documentation.
• 80% Cost Reduction — Case study achieving 80% infrastructure cost reduction through tuning.
• Memory Architecture Guide — Developer guide for configuring and scaling OpenClaw's memory architecture.
• Memory, Concurrency & Context — Practical guide to optimizing memory, concurrency, and context handling.
• Multi-Agent Coordination — Patterns for coordinating multiple OpenClaw instances in multi-agent workflows.
• 4 Weeks in Production — Real-world lessons from running OpenClaw in production for four weeks.
• Cost Calculator — Real-time cost estimation tool.
• Cost Governor — Budget enforcement tool for OpenClaw usage.
• clawdbot-cost-monitor — Open-source cost monitoring dashboard.

• Official CLI Backup — Official CLI documentation for backup and restore operations.
• Complete Backup Guide — Comprehensive guide covering full backup and restore procedures.
• Data Protection Guide — End-to-end data protection guide for OpenClaw deployments.
• Settings & Memory Export — Tutorial for exporting OpenClaw settings and memory.
• Disaster Recovery on Hetzner — Disaster recovery setup with restore drills and practical failover on Hetzner.
• openclaw-backup Tool — Open-source CLI tool for automated OpenClaw backups.
• Keep My Claw — Managed zero-knowledge encrypted backup service for OpenClaw.

• Enterprise Multi-User & Hardening — Guide for multi-user enterprise deployments with hardening applied.
• Enterprise Security Setup — Security-focused enterprise setup walkthrough.
• Enterprise Automation Use Cases — Survey of enterprise automation use cases and implementation patterns.
• Customer Service to Sales — Case studies ranging from customer service automation to sales workflows.
• Business Workflows & Guardrails — Case studies examining where autonomous AI creates value and where guardrails are needed.
• One Month Review — Honest one-month post-deployment review of enterprise OpenClaw usage.
• Presidio: NVIDIA Ecosystem — Enterprise deployment in the NVIDIA ecosystem with Presidio's perspective.
• Implementation Guide — Practical implementation guide for business teams adopting OpenClaw.
• DataStudios: Enterprise Case Studies — Real-world deployments at TELUS (57,000 employees), Zapier (10× growth), and Tines (100× acceleration).
• Faros AI: Claude Code ROI — Measured 4:1 ROI with $37.50 per incremental PR vs $150 saved developer time.

• RBAC Feature Request #8081 — Upstream issue tracking role-based access control support.
• Agents Plane #17299 — Upstream issue tracking multi-tenant agent plane support.
• openclaw-multitenant — Container isolation with encrypted vault and team sharing support.
• lobu — Multi-tenant OpenClaw solution designed for team deployments.
• Session Isolation Vulnerability — Research exposing session isolation failures, authorization bypass, and privilege escalation in multi-user setups.

• VPS Monitoring — Tutorial for uptime, log, metrics, and alert monitoring on VPS deployments.
• SigNoz: OpenTelemetry Integration — Official SigNoz documentation for OpenClaw OpenTelemetry integration.
• SigNoz: Dashboard Guide — Guide for building OpenClaw monitoring dashboards with OpenTelemetry.
• Comet/Opik: Native Observability — Native observability integration using Comet and Opik.
• openclaw-observability-plugin — Open-source observability plugin for OpenClaw.
• ClawMetry — Real-time metrics dashboard for OpenClaw instances.
• BytePlus: Trace Monitoring — Distributed trace monitoring for OpenClaw via BytePlus Observability Platform.
• Grafana Claude Stats Plugin — Official Grafana plugin for real-time OpenClaw statistics dashboard.
• Grafana Cloud Anthropic Integration — Monitor Claude usage and costs with Grafana Cloud.
• Mission Control — War-room style monitoring with RBAC.
• Hawk Eye — Workspace sentinel for anomaly behavior detection.
• OpenClaw Studio — Visual agent management tool.
• ClawController — Real-time monitoring dashboard.

• SOC 2 / HIPAA / GDPR Guide — Enterprise compliance guide covering SOC 2, HIPAA, and GDPR requirements.
• 2026 Enterprise Hardening — Up-to-date enterprise hardening guidance for 2026.
• CLAW-10 Evaluation Framework — Structured framework for evaluating OpenClaw enterprise readiness.
• SOC 2 Audit Log Gap — Important finding: OpenClaw lacks the audit logging required for SOC 2 compliance.
• Security Model Deep Dive — Comprehensive deep dive into OpenClaw's underlying security model.
• Official SECURITY.md — Official security policy and vulnerability reporting instructions.

• Salesforce MCP — Salesforce integration via MCP for OpenClaw agents.
• Salesforce Service Cloud — Salesforce Service Cloud integration for support automation.
• Jira MCP — Jira integration via MCP for project management automation.
• ServiceNow MCP — ServiceNow integration for IT service management workflows.
• Jira Integration Guide — Step-by-step guide for connecting OpenClaw to Jira.
• Expanso: Jira Automation — Expanso's guide for automating Jira workflows with OpenClaw.
• ServiceNow Agent Skill — Community-built ServiceNow agent skill for ClawHub.

• SecureClaw (Adversa AI) — First OWASP ASI Top 10 aligned security plugin with MITRE ATLAS mapping, dual-layer defense (code + behavior).
• ClawSecure — 3-layer audit protocol detecting 55+ OpenClaw-specific threat patterns; audited 2,890+ agents.
• clawscan — Official ClawHub static scanner skill for pre-install security checks.
• openclaw-security-monitor — Active monitoring for ClawHavoc, AMOS Stealer, CVE-2026-25253, and supply chain attacks.
• OWASP Agentic Skills Top 10 — Official OWASP checklist for agent skill security.
• arc-security-audit — Full agent skill stack security audit.
• clawdefender — Security scanner + input sanitizer skill.
• agentkeys — Secure credential broker for AI agents.

• Salesforce MCP — Natural language CRM automation via Composio.
• Jira Integration — Issue creation, sprint tracking, JQL queries, workflow automation.
• sql-toolkit — PostgreSQL/MySQL/SQLite with NL-to-SQL, parameterized queries, read-only mode, audit logging.
• database — Unified PostgreSQL/MySQL/SQLite/MongoDB/Redis access with SSH tunnel and destructive operation confirmation.
• GOG (Google Workspace) — Gmail + Calendar + Drive + Sheets integration, 14,000+ installs.

• openclaw-operator — Production Kubernetes operator with network isolation, secret management, and health monitoring.
• openclaw-dashboard — Secure real-time monitoring panel with Auth, TOTP MFA, cost tracking, and memory browser.
• GitHub Actions CI/CD Monitoring — Detect build/test failures and auto-pull error logs.
• Lobster Shell — Native OpenClaw workflow shell composing skills and tools into reusable pipelines.

• Semgrep Security Cheat Sheet — Static analysis rules for skill security auditing.
• OpenClaw Security Best Practices — Pre-install audit steps and minimal sandbox permissions.
• Official Security Docs — Permission model, sandbox config, --allow-tools deny-by-default.
• Snyk ToxicSkills Report — Audit of 22,511 skills: 36% with defects, 1,467 malicious payloads, 127 invisible Unicode steganography.
• VirusTotal Integration — All ClawHub uploads now scanned via VirusTotal Code Insight.

• edict — 9-agent orchestration system inspired by ancient Chinese "Three Departments and Six Ministries" governance. Real-time dashboard, kanban view, full audit trails, multi-model support (13.4K stars). NEW
• openclaw-a2a-gateway — A2A (Agent-to-Agent) protocol v0.3.0 implementation — bidirectional agent communication gateway for standardized multi-agent interoperability. NEW
• agent-team-orchestration — Role definitions, task lifecycle, and handoff protocols.
• agent-collaboration-network — Agent registry, skill-based discovery, message routing.
• arc-department-manager — Manage AI sub-agent departments.

• Official Skills Guide — Creating custom skills with SKILL.md.
• Plugin SDK — TypeScript SDK for channels, model providers, tools, and skills.
• DataCamp Tutorial — Hands-on custom skill building guide.
• DigitalOcean Guide — Comprehensive skill overview and development walkthrough.
• VoltAgent/awesome-openclaw-skills — 5,400+ curated skills from the official registry.

Memory is one of the fastest-evolving areas in the OpenClaw ecosystem. Two complementary approaches have emerged: long-term memory (cross-session knowledge) and context management (within-session lossless handling).

Best practice: Enable one long-term memory plugin + one context management plugin together.

• memory-lancedb-pro — Hybrid retrieval long-term memory — Vector+BM25+Cross-Encoder reranking, auto-captures 6 knowledge categories, Weibull decay model for intelligent forgetting (3.7K stars). NEW
• memU — Built for 24/7 persistent agents — supports Claude Skills + MCP protocol, designed for always-on agentic workflows (13.2K stars). NEW
• Mem0 Integration — Most popular persistent memory solution for OpenClaw agents.
• OceanBase PowerMem — Powerful memory backend by Alibaba OceanBase.
• Supermemory — Supermemory integration for OpenClaw.
• MoltBrain — Brain-like long-term memory system.
• Engram — Unified multi-agent memory framework.
• openclaw-memory-mem0 — Semantic fact extraction memory layer.
• claude-mem — Claude-specific memory layer.

• lossless-claw — Lossless Context Management — DAG-based hierarchical summarization + SQLite persistence, zero information loss. Agents can search (lcm_grep) and drill into (lcm_expand) any compacted history (3.7K stars). NEW
• OpenViking — ByteDance's context database for AI agents — file-system paradigm for unified management of memory, resources, and skills. Supports hierarchical context delivery and self-evolution (19.8K stars). NEW
• MemOS — Memory operating system for AI agents.
• MemOS Cloud Plugin — Cloud-based memory plugin for OpenClaw.
• memov — Universal memory layer and retrieval toolkit.
• openamnesia — Continuous learning context engine with privacy-safe memory extraction.
• Cognee — Graph-based memory and recall system.

A brief overview of notable OpenClaw forks and related frameworks.

Contributions are welcome. Please:

1. Check that the link is alive and directly relevant to OpenClaw security, self-hosting, or enterprise use.
2. Add a short, factual description (one sentence).
3. Place the entry in the most appropriate section.
4. Submit a pull request with a clear title.

To the extent possible under law, the contributors have waived all copyright and related rights to this work.