인프라·DevOps 스킬

Assemble a reproducible Alchemy incident bundle with sanitized environment details, request IDs, timing, status evidence, and reproduction results. It helps teams prepare support evidence while screening out credentials and sensitive customer data before sharing.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Preparing evidence for an Alchemy support case
• Investigating intermittent Ethereum RPC timeouts
• Reviewing a diagnostic bundle for leaked secrets

Collect evidence for reproducible Algolia search or indexing failures, including client versions, reproduction steps, and redacted request details. Package the findings with a file manifest and secret scan results for review before a support handoff.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Preparing evidence for an Algolia support request
• Documenting a reproducible search or indexing failure
• Reviewing diagnostic files for sensitive information

Design Alchemy traffic controls from observed account usage and workload patterns. This skill helps set shared admission limits, bounded retries, and load tests so teams can handle rate limits, protect critical requests, and define clear degraded behavior.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Handling repeated Alchemy 429 responses
• Protecting critical reads during traffic spikes
• Planning shared limits across workers

Review an Adobe integration before launch or a major change. This skill checks authentication, entitlements, API versions, data handling, asynchronous reliability, spending, monitoring, and rollback, then records a go, conditional go, or no-go decision with evidence.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing an Adobe integration before launch
• Checking permissions and API versions
• Assessing data handling and async job recovery

Plan and verify separation between Adobe development, staging, and production environments. Produce an environment matrix, configuration manifest, isolation tests, and release controls covering credentials, storage, events, budgets, and ownership before making live changes.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Design separate Adobe development and production environments
• Verify staging cannot access production storage
• Detect workspace and product profile mismatches

Triage Adobe integration incidents across authentication, products, jobs, events, storage, spending, and App Builder. It helps classify the failure, preserve evidence, contain impact, test recovery with a bounded canary, and record customer impact and follow-up actions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Investigating an Adobe authentication or credential incident
• Containing a Firefly outage or failed async jobs
• Reviewing event delivery, storage, or spending issues

Plan controlled load tests for Adobe services and establish a conservative operating capacity. Measure queues, completion rates, errors, costs, and output correctness, then document scaling controls, emergency stop conditions, and evidence specific to the tested environment.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Planning capacity for Adobe processing workloads
• Checking queue behavior as load increases
• Testing recovery after worker restarts

Deploy Adobe-backed services or App Builder apps with separate environments, verified build artifacts, and limited trial traffic. The workflow checks monitoring and rollback readiness, then records evidence for release decisions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Deploying to Stage while verifying Production stays unchanged
• Evaluating a release with limited trial traffic
• Testing rollback after a simulated entitlement failure

Design service-specific controls for Adobe workloads, including request admission, Retry-After handling, capped retries, and queues with spend limits. It helps respond to 429 errors and scaling pressure using verified service constraints and test evidence.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Handling Adobe 429 responses using Retry-After
• Setting bounded retries for transient idempotent work
• Planning queue limits and backpressure for larger workloads

Plan capacity, queues, and retry rules for an Abridge integration using tenant-specific contracts and verified responses. Define how to handle overload, prevent duplicate clinical actions, and hand off work when processing cannot continue.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Plan processing capacity for peak Abridge workloads
• Define retry rules that prevent duplicate clinical actions
• Set queue limits and procedures for stale work

Review Adobe integrations for approved endpoints, versions, data handling, spending limits, and approval requirements. Produce clear allow and deny rules, enforcement points, tests, and evidence so teams can check risky actions before they run.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing Adobe API calls before enabling them
• Blocking obsolete routes or unknown hosts
• Checking logs for leaked tokens or signed URLs

Build and deploy Google ADK agents to Vertex AI Agent Engine. Define sequential, parallel, or looping workflows and A2A interfaces for agent communication, then verify endpoints and add logging and retry behavior.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Deploying ADK agents to Vertex AI Agent Engine
• Designing sequential, parallel, or looping agent workflows
• Verifying AgentCard discovery and task submission APIs

Build and validate Terraform for Vertex AI ADK infrastructure, including Agent Engine, networking, IAM permissions, code execution, and Memory Bank. It helps produce deployment settings and a checklist for security requirements such as VPC Service Controls.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Provisioning a Vertex AI ADK agent runtime
• Setting up secure VPC networking and IAM
• Adding code execution and Memory Bank

Design recovery controls for asynchronous Adobe workloads, including duplicate prevention, time-limited status checks, service-specific queues, and failure isolation. Produce state models and recovery runbooks that help teams investigate uncertain jobs before approving a replay.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Plan recovery after a submission-time crash
• Investigate an unknown job before replay
• Design service isolation and outage recovery drills

Design monitoring for Adobe integrations without logging secrets or content. Define structured logs, connect requests to jobs, establish service health measures, and prepare alerts, response procedures, and evidence retention checks.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Trace asynchronous jobs without recording prompts or output URLs
• Review monitoring gaps across App Builder execution modes
• Design actionable alerts for 429 errors and disabled events

Prepare an Abridge implementation by mapping tenant access, identity owners, roles, and EHR integration requirements to approved evidence. The result is a readiness inventory with verified facts, unresolved gaps, and actions that still need approval.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Preparing access and ownership details for an Abridge onboarding
• Checking identity roles and least-privilege provisioning requirements
• Documenting EHR integration prerequisites from tenant-specific evidence

Choose an Adobe authentication flow based on who owns the data, confirm the required entitlement and product profile, and document secret handling. The workflow yields a reviewable setup plan and a read-only proof before integration work proceeds.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing authentication for organization-owned data
• Setting up user consent for user-owned data
• Checking Adobe product entitlement and scopes

Use KubeKey’s `kk` commands to create and configure Kubernetes clusters, add or remove nodes, renew certificates, and manage offline installation artifacts. The skill explains inventory and config files, common workflows, prechecks, and troubleshooting guidance.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Creating a Kubernetes cluster with KubeKey
• Adding or removing cluster nodes
• Preparing offline Kubernetes installation artifacts

Turn authorized external reconnaissance into ongoing monitoring. Track changes in assets and findings across scans, watch public sources for adversary chatter, manage finding states, and send threshold-based webhook alerts with retry guidance and scheduled-run examples.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Compare this week's perimeter with the previous scan
• Set up recurring exposure monitoring
• Track ransomware and brand chatter

Prepare and review software releases by checking the Release Please pull request, version, changelog, and draft files. It guides verification and the manual publishing workflow, keeping release publication tied to explicit authorization.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing a proposed version and changelog
• Checking draft release assets before publication
• Preparing a release after changes reach main

Prepare OpenEnv environments for deployment to Hugging Face Spaces, check that the private Spaces work at runtime, and manage versioned collections and canonical references. It guides release decisions using build and usability checks.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Deploy selected OpenEnv environments to private Hugging Face Spaces
• Check runtime behavior before promoting a release
• Add successful Spaces to a versioned collection

Deploy an OpenEnv environment to Hugging Face Spaces with the OpenEnv CLI. Configure options such as visibility, web interface, base image, or hardware, then report the Space URL and check deployment health after the build completes.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Publishing an OpenEnv environment to Hugging Face Spaces
• Updating an existing Space after environment changes
• Deploying a private Space or API-only environment

Diagnose OpenEnv deployments on Hugging Face Spaces, identify build or runtime problems, and guide targeted recovery. It also covers checking whether a Space is healthy and reconciling release collection membership after redeployments.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Investigating a Space stuck while building or starting
• Finding the cause of a runtime or build error
• Redeploying one affected OpenEnv environment

Prepare Benny’s configuration for Slack-based bug triage and issue reproduction workflows. It guides setup of repository, tracker, routing, control, model, and budget choices, then checks the integrations and safeguards needed before creating or updating Cursor automations.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Setting up Benny for a project
• Choosing Slack and issue tracker settings
• Preparing triage and reproduction automations