families / devops
devops
app-store-connect-automation
A suite of skills for managing the Apple App Store Connect lifecycle, including metadata, builds, releases, and distribution.
These skills focus exclusively on Apple App Store Connect (ASC) workflows and iOS/macOS distribution, distinct from general DevOps tools like Python deployment or App Insights instrumentation.
- Automate the upload of new app builds and sync metadata to App Store Connect using the CLI.
- Configure ad-hoc distribution profiles and handle notarization for a new iOS application release.
aws-infrastructure-management
A collection of skills for provisioning, configuring, and managing core AWS infrastructure services including compute, storage, networking, and databases.
These skills focus on the direct configuration and lifecycle management of specific AWS resources and services, distinct from cost optimization strategies or initial context discovery.
- Set up a VPC with public and private subnets and deploy an EC2 instance with a security group.
- Configure an S3 bucket with lifecycle rules and create an RDS database with automated backups.
azure-infrastructure-management
A collection of skills for provisioning, configuring, and managing core Azure infrastructure components including compute, networking, databases, and serverless functions.
These skills focus on the underlying infrastructure and resource provisioning (compute, network, data, serverless) rather than CI/CD pipeline orchestration or specific application onboarding workflows.
- Deploy a hub-spoke network topology with NSGs and private endpoints using ARM templates.
- Provision an Azure SQL Database with automated backups and configure an Azure Function to trigger on new data.
maven-build-optimization
A set of skills for configuring, structuring, and optimizing Apache Maven build processes to improve performance and reliability.
These skills focus specifically on the mechanics, configuration, and performance tuning of the Maven build system, distinct from broader platform engineering or unrelated data processing tasks.
- Configure Maven plugins to enable parallel execution and reduce overall build time.
- Analyze the Maven build lifecycle to identify bottlenecks and establish a performance baseline.
ci-pipeline-management
A set of skills for designing, implementing, and optimizing continuous integration pipelines across various cloud and local environments.
These skills differ by their specific focus on either authoring workflow syntax, optimizing build performance through caching, enforcing structural contracts, or managing the broader deployment engineering lifecycle.
- Configure a GitHub Actions workflow to cache Playwright dependencies and enforce pipeline contracts.
- Analyze the current CI build times and set up a new cloud build pipeline for faster deployment.
cloud-architecture-design
A set of skills for discovering, visualizing, and defining high-level cloud infrastructure and service architectures.
These skills focus on the structural design, discovery, and visualization of cloud systems, distinct from operational management, cost optimization, or specific provider tooling.
- Create a holistic view of our current cloud architecture to identify single points of failure.
- Generate a cloud architecture canvas for a new microservices deployment on GCP.
container-image-management
A set of skills for building, managing, and deploying container images and their associated registries.
These skills differ by focusing on specific stages of the container lifecycle, such as registry administration, local development, or orchestration deployment, rather than a single unified workflow.
- Push a new Docker image to Amazon ECR and configure the repository retention policy.
- Build a local Docker image for a microservice and deploy it to a Kubernetes cluster.
data-backup-and-ingestion
A set of skills for managing data persistence through various backup mechanisms and executing Monte Carlo push ingestion workflows.
The skills differ by their specific implementation targets, ranging from generic or named backup instances (like moltbook or specific openclaw variants) to the distinct data ingestion process of monte-carlo-push-ingestion.
- Execute the openclaw-backup routine to snapshot the current system state.
- Run the monte-carlo-push-ingestion job to process the latest data batch.
database-operations
A collection of skills for managing database schemas, executing migrations, handling backups, and performing administrative tasks across various database systems.
These skills differ by their specific focus on schema evolution and migration execution versus backup strategy design and general administrative operations.
- Execute a zero-downtime migration to add a new column to the users table using Drizzle.
- Configure automated daily backups with a 30-day retention policy and test the recovery process.
deployment-operations
A family of skills covering the execution, validation, and management of software releases across various environments and platforms.
These skills differ by their specific focus on either the architectural design of pipelines, the execution of specific platform deployments, or the operational management of release states such as promotion and rollback.
- Execute a production deployment using the standard devops-deploy skill and verify the service health.
- Design a multi-stage CI/CD pipeline with approval gates for a new microservice.
environment-setup
A collection of skills for configuring, installing, and validating software development and runtime environments across various platforms and frameworks.
These skills differ by the specific software stack (e.g., TAO, HoloScan, Earth2Studio) or the target execution platform (e.g., Docker, Kubernetes, SLURM) they configure or verify.
- Set up the NVIDIA TAO toolkit environment on a local GPU host.
- Check if the current Python environment has the required dependencies for the project.
terraform-cloud-provisioning
Provision and manage infrastructure across AWS, Azure, and GCP using Terraform or OpenTofu, including module creation, state management, and migration strategies.
This cluster focuses specifically on Infrastructure as Code (IaC) tools like Terraform and OpenTofu for cloud resource provisioning, distinguishing it from general Git workflows, AI pipeline orchestration, or service mesh management.
- Create a reusable Terraform module for an AWS VPC with subnets and security groups.
- Migrate an existing Terraform project to OpenTofu and update the CI/CD pipeline accordingly.
gke-tpu-ai-troubleshooting
Diagnose and resolve specific hardware and software issues related to TPU and MXLA workloads on GKE, including hangs, performance degradation, and resource errors.
This skill focuses exclusively on troubleshooting TPU and MXLA-specific failures and performance issues, whereas other GKE skills cover general cluster management, networking, or non-AI workloads.
- Investigate why my TPU training job is hanging during the initialization phase.
- Analyze the cause of sudden performance degradation in my MXLA inference service.
incident-response
A family of skills covering the full lifecycle of production incident management, from real-time triage and mitigation to post-incident analysis and process improvement.
These skills differ by their focus on specific phases of the incident lifecycle, such as real-time response and mitigation versus post-incident documentation and process configuration.
- Coordinate the triage and mitigation of a critical production outage and document the timeline for the postmortem.
- Create a structured runbook for database failure scenarios and configure the on-call escalation paths.
infrastructure-deployment-and-management
A set of skills for designing, deploying, and validating infrastructure using code, cloud platforms, and specialized hardware.
These skills differ by their specific focus on either general infrastructure-as-code patterns, specific cloud provider operations like Azure, or specialized hardware deployments such as Doca and physical AI infrastructure.
- Write a Terraform module to deploy a resilient Azure Kubernetes cluster and validate the configuration.
- Deploy a Doca BF3 smart NIC configuration on a bare-metal server and verify the interconnect status.
jetson-hardware-customization
A set of skills for configuring and modifying the hardware interfaces, power management, and peripheral settings of NVIDIA Jetson modules.
These skills focus on low-level hardware configuration and peripheral customization, distinct from image flashing, BSP downloading, or general diagnostic tasks.
- Configure the pinmux settings to enable a specific GPIO header for a custom sensor.
- Adjust the nvpmodel and fan speed profiles to optimize thermal performance for a high-load AI workload.
k8s-ai-workload-management
A set of skills for deploying, scaling, and troubleshooting AI and GPU/TPU workloads on Kubernetes and GKE.
The skills differ by focusing on either the proactive creation and scaling of AI workloads (manifests, autoscaling) or the reactive diagnosis of specific infrastructure and job failures (GKE troubleshooting, TPU monitoring).
- Create a production-ready Kubernetes manifest for an LLM inference service with GPU resource limits.
- Troubleshoot a GKE jobset interruption caused by node unresponsiveness during a TPU training run.
kubernetes-lifecycle-management
A family of skills covering the design, deployment, configuration, and operational management of Kubernetes workloads and clusters.
These skills differ in their focus, ranging from high-level architectural design and security auditing to specific operational tasks like workload deployment, scaling, and chart building.
- Create a Helm chart for a microservice and deploy it to a production cluster with autoscaling enabled.
- Troubleshoot a crashing pod by analyzing logs and resource limits, then update the deployment configuration to fix the issue.
llm-inference-optimization
Optimize LLM inference performance and cost through caching strategies, gateway management, and traffic routing.
This cluster focuses on runtime inference efficiency, cost reduction, and traffic management, distinct from model training (fine-tuning) or observability (langsmith, phoenix).
- Implement a multi-layer caching system using Redis and semantic similarity to reduce LLM API latency and costs.
- Deploy an LLM gateway with LiteLLM Proxy to handle load balancing, rate limiting, and fallback routing for multiple providers.
model-lifecycle-management
A family of skills covering the governance, deployment, and operational orchestration of machine learning models within enterprise infrastructure.
These skills differ by their focus on the specific phase of the model lifecycle, ranging from high-level governance and registry standards to low-level infrastructure deployment and serving orchestration.
- Define the metadata schema and approval workflow for new model versions in our enterprise registry.
- Deploy a new LLM to Kubernetes using KServe with GPU autoscaling and canary release strategy.
cloud-monitoring-operations
A set of skills for configuring, querying, and visualizing metrics within cloud monitoring platforms.
These skills focus specifically on the configuration, querying, and visualization of metrics within cloud monitoring ecosystems, distinct from general ops monitoring or specific hardware telemetry exporters.
- Generate a PromQL query to identify the top 5 services by latency and create a corresponding chart.
- Configure a notification channel to alert the on-call team when CPU usage exceeds 90% for more than 5 minutes.
network-infrastructure-management
A collection of skills for managing, auditing, and optimizing network hardware, IP resources, and data center connectivity.
These skills differ by their specific focus on either general network auditing, vendor-specific device management (Arista, Mikrotik), or specialized data center hardware acceleration (Doca).
- Audit the IPAM and DNS records to identify orphaned entries and conflicts.
- Check the health status of Arista switches and configure RDMA offloading on Doca devices.
nvidia-doca-networking
A suite of skills for configuring, optimizing, and troubleshooting NVIDIA DOCA networking components including flow processing, DPA acceleration, and Ethernet interfaces.
These skills focus specifically on the NVIDIA DOCA software stack for networking and data path acceleration, distinguishing them from general DevOps or other NVIDIA hardware management tasks.
- Configure a DOCA flow rule to redirect traffic from a specific IP to a DPA pipeline.
- Optimize the performance of the DOCA Ethernet driver using the flow-tune utility.
observability-implementation
A family of skills focused on designing, configuring, and instrumenting systems to provide visibility into application and infrastructure health through metrics, logs, and traces.
These skills differ by their specific focus on either high-level system design, general monitoring setup, or language-specific instrumentation and configuration.
- Set up distributed tracing and alerting for a new microservice deployment.
- Instrument a .NET application with OpenTelemetry to capture production performance metrics.
package-lifecycle-management
A set of skills for managing the installation, configuration, and distribution of software packages across various environments and platforms.
These skills differ by their specific focus on either the source of the package (mirrors, wheels, debs), the target environment (Debian, research repos), or the lifecycle stage (publishing, post-install configuration).
- Install the HoloScan wheel package on a Linux system and configure the post-installation settings.
- Resolve the correct China mirror for a package dependency and publish the updated package to the internal marketplace.
performance-testing-and-optimization
A suite of skills for executing load tests, diagnosing bottlenecks, and monitoring system performance to ensure optimal application speed and stability.
These skills differ by their specific focus on either generating load (k6, loadpage), diagnosing root causes (monte-carlo, detecting-regressions), or providing expert strategic guidance (performance-engineer, tune-monitor).
- Run a k6 load test against the API endpoint and generate a report on p95 latency.
- Analyze the recent performance metrics to detect any regressions in database query times.
sandbox-lifecycle-management
A set of skills for provisioning, configuring, and managing the lifecycle of isolated development and testing environments.
These skills differ by their specific role in the sandbox lifecycle, ranging from initial setup and simulation to version migration and post-copy configuration automation.
- Set up a new isolated sandbox environment for testing a specific microservice.
- Migrate the current sandbox configuration to the next stable release version.
service-mesh-management
Design, implement, and monitor service mesh infrastructure using Istio or Linkerd to manage microservice communication, security, and observability.
This cluster focuses specifically on the configuration, security, and observability of service mesh technologies like Istio and Linkerd, distinguishing it from general service lifecycle management or standalone monitoring tools.
- Configure mTLS and traffic routing policies for a multi-cluster Istio deployment.
- Set up distributed tracing and latency SLOs for a Linkerd service mesh.
system-health-and-integrity
A collection of skills for monitoring, diagnosing, and maintaining the operational health, time synchronization, and integrity of Linux and macOS systems.
These skills focus specifically on the ongoing health, time accuracy, and data integrity of the host system, distinct from deployment, migration, or specific application debugging tasks.
- Check the current system health and verify that the system time is synchronized with NTP servers.
- Run a comprehensive integrity check to ensure backups are valid and the system is free from corruption.
workflow-automation-orchestration
A set of skills for designing, building, debugging, and managing automated business processes across various orchestration platforms.
These skills differ by the specific automation platform they target (e.g., n8n, Power Automate, Zapier) and the stage of the workflow lifecycle they address (e.g., setup, building, debugging, or general management).
- Set up a self-hosted n8n instance and create a workflow that triggers on a new Slack message.
- Diagnose why a Power Automate flow is failing at the data mapping step and suggest a fix.