# Render tasks for coding agents

Use this document after setup, when the Render MCP server is connected and the Render skills are installed. For setup, read https://render.com/agents.md first.

Match the user request to one cluster below. Invoke the named skill, call the listed MCP tools, and read only the linked docs pages.

Confirm the connection with `list_workspaces`, scope the session with `select_workspace`, and check the current scope with `get_selected_workspace`. Every service tool takes a service ID: resolve a service the user names informally with `list_services`, and read its configuration with `get_service`.

## Deploy a new app

Triggers: "deploy this", "ship this to production", "host my app on Render".
Skill: `render-deploy`. Add `render-blueprints` when the repo needs a `render.yaml`, or `render-static-sites` for a static build.
MCP: `create_web_service`, `create_static_site`, or `create_cron_job`, then `trigger_deploy` and `get_deploy` to follow the result.
Docs: https://render.com/docs/your-first-deploy.md, https://render.com/docs/web-services.md, https://render.com/docs/blueprint-spec.md

## Debug a failed deploy or a failing service

Triggers: "my build is failing", "the deploy failed", "it returns 502", "the service will not start".
Skill: `render-debug`.
MCP: `list_deploys`, then `get_deploy` on the failed deploy. Read `list_logs` for errors, narrowing the filters with `list_log_label_values`, and `get_metrics` for CPU and memory.
Docs: https://render.com/docs/troubleshooting-deploys.md, https://render.com/docs/deploys.md, https://render.com/docs/health-checks.md

## Add or manage a Postgres database

Triggers: "add a database", "connect Postgres", "query production data".
Skill: `render-postgres`.
MCP: `create_postgres`, `list_postgres_instances`, `get_postgres`, and `query_render_postgres` for read-only SQL.
Docs: https://render.com/docs/postgresql-creating-connecting.md, https://render.com/docs/postgresql-connection-pooling.md, https://render.com/docs/postgresql-backups.md

## Add or manage a Key Value store

Triggers: "add Redis", "I need a cache", "session store", "queue backend".
Skill: `render-keyvalue`.
MCP: `create_key_value`, `list_key_value`, `get_key_value`.
Docs: https://render.com/docs/key-value.md, https://render.com/docs/valkey-faq.md

## Environment variables and secrets

Triggers: "set an env var", "DATABASE_URL is missing", "add my API key".
Skill: `render-env-vars`.
MCP: `update_environment_variables`. It merges with the existing variables on the service by default, so send only the ones that changed; set its `replace` parameter to true only when you need to drop variables.
Docs: https://render.com/docs/configure-environment-variables.md, https://render.com/docs/environment-variables.md

## Custom domains and TLS

Triggers: "point my domain at this", "add HTTPS", "the certificate is not issuing".
Skill: `render-domains`.
MCP: `get_service` to read the current hostnames. There is no domain or certificate tool; make those changes in a Blueprint, the Render API, or the dashboard.
Docs: https://render.com/docs/custom-domains.md, https://render.com/docs/tls.md, https://render.com/docs/configure-other-dns.md

## Scaling and compute plans

Triggers: "handle more traffic", "the service is slow under load", "reduce my bill".
Skill: `render-scaling`.
MCP: `get_metrics` for CPU, memory, and instance counts to size the change. `update_web_service` does not apply it and only returns a dashboard link, so make the change in a Blueprint, the Render API, or the dashboard.
Docs: https://render.com/docs/scaling.md, https://render.com/docs/compute-plans.md

## Run untrusted code in a sandbox

Triggers: "run this code in isolation", "execute user-submitted code", "give my agent a place to run code".
Skill: `render-sandboxes`.
MCP: There is no sandbox tool. Create sandboxes and run commands with the Render SDK or CLI, as the skill describes.
Docs: https://render.com/docs/sandboxes.md, https://render.com/docs/sandboxes-creating.md, https://render.com/docs/sandboxes-executing.md, https://render.com/docs/sandboxes-files.md

## Logs, metrics, and alerts

Triggers: "show me the logs", "why is CPU high", "tell me when it goes down".
Skill: `render-monitor`.
MCP: `list_logs`, `list_log_label_values` to discover the filters worth using, and `get_metrics` for services and datastores.
Docs: https://render.com/docs/logging.md, https://render.com/docs/service-metrics.md, https://render.com/docs/notifications.md

## Anything else

Setup, authentication, and skill installation: https://render.com/agents.md
Search the docs: https://render.com/api/agent/docs-search?query=
Read any docs page as markdown: https://render.com/docs/{slug}.md