Openclaw Security Hardening Agent Skill
Harden OpenClaw self-hosted environments with baseline host controls, auth tightening, secret handling, network segmentation, and safe update/rollback workflows. Use when deploying OpenClaw in home labs, startups, or production-like local AI infrastructure.
Install
one command, takes just this skill from the repositorynpx skills add https://github.com/BagelHole/DevOps-Security-Agent-Skills --skill openclaw-security-hardeningThe instruction itself
9 sections, as written by the authorOpenClaw Security Hardening
Use this skill to reduce exposure in self-hosted OpenClaw deployments before opening access to teammates or external networks.
Build a Threat Model First
Map the highest-risk assets and paths:
- Admin/API endpoints for OpenClaw
- Provider API keys and model credentials
- Prompt/response logs containing sensitive business data
- Host-level access (SSH, local admin accounts, remote desktop)
Prioritize controls that reduce credential theft, remote code execution blast radius, and data exfiltration.
Apply Baseline Host Hardening
- Keep OS and package dependencies patched on a regular cadence.
- Run OpenClaw as a dedicated non-admin user account.
- Enable full-disk encryption and secure boot features where available.
- Remove unnecessary services and block inbound ports by default.
- Lock down remote admin (key-only SSH, no password login, limited source CIDRs).
Example Linux baseline checks:
id openclaw
sudo ss -tulpn
sudo ufw status verbose
sudo systemctl --failed
Harden Application Runtime
- Bind OpenClaw to localhost or private VLAN by default.
- Place a reverse proxy in front of OpenClaw for TLS, auth, and rate limits.
- Enforce authentication on every non-health endpoint.
- Disable debug/dev modes in persistent environments.
- Restrict outbound egress to only required providers (LLM API, telemetry sink, package mirror).
Example reverse proxy controls to enforce:
- TLS 1.2+ only
- strict transport security header
- request body size limits
- request timeout and upstream timeout guardrails
- per-IP and per-token rate limiting
Protect Secrets and Tokens
- Store secrets in a vault or platform secret manager, not committed
.envfiles. - Rotate provider and admin tokens on a fixed interval and after any incident.
- Scope tokens minimally (least privilege, per-service keys).
- Scan repos and deployment artifacts for leaked credentials before release.
Rotation checklist:
- Generate replacement key.
- Update runtime secret store.
- Restart or reload OpenClaw.
- Validate request success with new key.
- Revoke old key.
Segment Network Access
Use layered access patterns:
- Tier 1 (private): OpenClaw service port reachable only from app/proxy subnet.
- Tier 2 (operator): Admin plane reachable only from VPN/Tailscale/WireGuard.
- Tier 3 (public): Expose only hardened reverse proxy with strict ACLs.
Do not publish raw OpenClaw service ports directly to the internet.
Add Detection and Recovery Paths
- Centralize auth, error, and audit logs.
- Alert on brute-force attempts, token failures, and unusual outbound traffic.
- Capture immutable backup snapshots of configs and prompt data retention settings.
- Test rollback and restore procedures every release cycle.
Minimum operational runbook:
- service restart path
- key revocation path
- incident isolation path (network block + token disable)
- known-good rollback version
Validation Checklist
- All sensitive endpoints require auth and are unreachable without VPN or gateway policy.
- Secrets are absent from repo history and plaintext shared directories.
- Host firewall default deny is active for inbound traffic.
- TLS termination and rate limits are active at ingress.
- Rollback drill can restore service within target RTO.
Related Skills
- openclaw-local-mac-mini - Local OpenClaw hosting setup
- multi-tenant-llm-hosting - Multi-tenant AI isolation patterns
- zero-trust - Private access and identity-aware network controls
Other skills for the same job
different authors, same section of the catalogue
Assess Kubernetes workloads and cluster configuration for AKS Automatic compatibility. Identifies incompatibilities, generates fixes, and guides migration from AKS Standard to AKS Automatic. WHEN: migrate to AKS Automatic, check AKS Automatic readiness, validate manifests for Automatic, assess cluster for Automatic compatibility, fix deployment for Automatic compatibility, identify AKS Automatic migration blockers, is my cluster ready for AKS Automatic.
Discovers available Azure OpenAI model capacity across regions and projects. Analyzes quota limits, compares availability, and recommends optimal deployment locations based on capacity requirements. USE FOR: find capacity, check quota, where can I deploy, capacity discovery, best region for capacity, multi-project capacity search, quota analysis, model availability, region comparison, check TPM availability. DO NOT USE FOR: actual deployment (hand off to preset or customize after discovery), quota increase requests (direct user to Azure Portal), listing existing deployments.
Interactive guided deployment flow for Azure OpenAI models with full customization control. Step-by-step selection of model version, SKU (GlobalStandard/Standard/ProvisionedManaged), capacity, RAI policy (content filter), and advanced options (dynamic quota, priority processing, spillover). USE FOR: custom deployment, customize model deployment, choose version, select SKU, set capacity, configure content filter, RAI policy, deployment options, detailed deployment, advanced deployment, PTU deployment, provisioned throughput. DO NOT USE FOR: quick deployment to optimal region (use preset).
Unified Azure OpenAI model deployment skill with intelligent intent-based routing. Handles quick preset deployments, fully customized deployments (version/SKU/capacity/RAI policy), and capacity discovery across regions and projects. USE FOR: deploy model, deploy gpt, create deployment, model deployment, deploy openai model, set up model, provision model, find capacity, check model availability, where can I deploy, best region for model, capacity analysis. DO NOT USE FOR: listing existing deployments (use foundry_models_deployments_list MCP tool), deleting deployments, agent creation (use agent/create), project creation (use project/create).
Intelligently deploys Azure OpenAI models to optimal regions by analyzing capacity across all available regions. Automatically checks current region first and shows alternatives if needed. USE FOR: quick deployment, optimal region, best region, automatic region selection, fast setup, multi-region capacity check, high availability deployment, deploy to best location. DO NOT USE FOR: custom SKU selection (use customize), specific version selection (use customize), custom capacity configuration (use customize), PTU deployments (use customize).
This skill should be used when working with LaminDB, an open-source data framework for biology that makes data queryable, traceable, reproducible, and FAIR. Use when managing biological datasets (scRNA-seq, spatial, flow cytometry, etc.), tracking computational workflows, curating and validating data with biological ontologies, building data lakehouses, or ensuring data lineage and reproducibility in biological research. Covers data management, annotation, ontologies (genes, cell types, diseases, tissues), schema validation, integrations with workflow managers (Nextflow, Snakemake) and MLOps platforms (W&B, MLflow), and deployment strategies.
Latch platform for bioinformatics workflows. Build pipelines with Latch SDK, @workflow/@task decorators, deploy serverless workflows, LatchFile/LatchDir, Nextflow/Snakemake integration.
Run Python code in the cloud with serverless containers, GPUs, and autoscaling. Use when deploying ML models, running batch processing jobs, scheduling compute-intensive tasks, or serving APIs that require GPU acceleration or dynamic scaling.
How to use it
Take bagelhole/openclaw-security-hardening from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.