Set up a Codex monitor for Vercel deployments and preview URLs. Use when the user invokes /vercel or $vercel, asks Codex to watch or monitor a Vercel deployment, waits for a Vercel preview or PR preview to become ready, or wants to be notified with both the deployment URL and preview URL once Vercel is READY.
npx skills add https://github.com/remotion-dev/remotion --skill vercel
Set up a short-lived Codex heartbeat that watches the remotion Vercel deployment and reports back in the current thread when it is ready or failed. Always include both the Vercel deployment/dashboard URL and the public preview URL in the final notification.
This repository has two Vercel deployments: bugs and remotion. The Vercel monitor skill always means the remotion deployment unless the user explicitly asks for another project.
remotion project.bugs deployment, even if it appears first or is newer.remotion. If necessary, inspect the candidates before creating a monitor.remotion.Prefer sources in this order:
remotion check or deployment status. Ignore the bugs check.remotion project, if the repository is linked and the user asked for the latest deployment.Use scripts/extract-vercel-links.py to normalize raw terminal, Vercel CLI, or GitHub output:
python3 .agents/skills/vercel/scripts/extract-vercel-links.py < deployment-output.txt
If only one link is available, use vercel inspect <url> when authenticated to confirm that it belongs to the remotion project and discover the missing deployment or preview URL. If neither a deployment URL nor a preview URL can be found, ask the user for one concise piece of context: the Vercel URL, PR URL/number, or branch name.
Prefer Vercel's deployment status over HTTP probing:
vercel inspect <deployment-or-preview-url>
Treat these as terminal states:
READY: report success immediately; do not create a monitor.ERROR, FAILED, CANCELED, or CANCELLED: report failure immediately; do not create a monitor.Treat these as monitorable states:
BUILDING, QUEUED, INITIALIZING, or unknown but plausibly in progress.If Vercel CLI is unavailable or unauthenticated, probe the preview URL with curl -I -L --max-time 20 <preview-url>. HTTP 2xx or 3xx means the preview is ready. HTTP 401 or 403 can also mean the deployment is ready but protected; report it as ready/protected if the response is clearly from Vercel deployment protection. A Vercel DEPLOYMENT_NOT_FOUND, 404, timeout, or DNS failure means keep monitoring unless a terminal Vercel status says otherwise.
Use the Codex app automation tool. If automation_update is not already in the tool list, search for it with tool_search before creating the monitor.
Create a heartbeat, not a cron, because the user wants this thread to be notified later. Use a one-minute cadence with a bounded count, usually 30 attempts unless the user asked for a different window. Do not show the raw RRULE string to the user.
The heartbeat prompt must be self-contained. Include:
vercel inspect <url>.Example heartbeat prompt:
Check this Vercel deployment until it reaches a terminal state.
Deployment URL: <deployment-url-or-unknown>
Preview URL: <preview-url-or-unknown>
Context: <branch/pr/project/commit-or-unknown>
Prefer `vercel inspect <deployment-or-preview-url>` if available and authenticated. If that is unavailable, use `curl -I -L --max-time 20 <preview-url>` and, if needed, fetch the response body to distinguish Vercel deployment protection from DEPLOYMENT_NOT_FOUND.
Ready means Vercel status READY, or the preview URL returns HTTP 2xx/3xx, or it returns Vercel deployment-protection HTTP 401/403. Failure means Vercel status ERROR, FAILED, CANCELED, or CANCELLED.
If ready, reply in the thread: "Vercel deployment is ready" and include:
- Deployment: <deployment-url>
- Preview: <preview-url>
If failed, reply in the thread with the failure status and include the same two links.
If still building, queued, initializing, not found, or unknown, stay quiet and let the next heartbeat check again. After reporting ready or failed, delete or pause this heartbeat if the automation id is available.
After creating the heartbeat, tell the user briefly which deployment/preview is being watched and the check cadence. If the deployment is already ready or failed, do not create a monitor; report the terminal status and links immediately.
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.
Take remotion-dev/vercel 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.