npx skills add ...
npx skills add contentstack/contentstack-agent-skills --skill cms-environments-publishing
npx skills add contentstack/contentstack-agent-skills --skill cms-environments-publishing
Advise developers on configuring environments, publishing content, using delivery and preview tokens, leveraging the Sync API, and understanding CDN and publish queue behavior in Contentstack.
Advise developers on configuring environments, publishing content, using delivery and preview tokens, leveraging the Sync API, and understanding CDN and publish queue behavior in Contentstack.
Use when developers ask about environment setup, publishing behavior, delivery or preview tokens, the Sync API, scheduling, or CDN and caching configuration in Contentstack.
Developers need to understand how content moves from draft to live, how environments and tokens work together, and how to keep frontends in sync.
Explains the publishing pipeline clearly. Identifies the correct token type for the scenario. Recommends the Sync API when appropriate. Avoids suggesting any unsafe client-side secret handling.
Understand the deployment pipeline. Recommend the environment structure. Explain token types and their use cases. Guide publishing, scheduling, and reference publishing. Recommend the Sync API for static sites or local caches.
[{"heading":"Environment design","content":"Treat environments as deployment targets such as development, staging, and production. Keep them aligned with the real deployment pipeline. Default max is 5 per stack. Environments are global modules and are shared across branches."},{"heading":"Publishing fundamentals","content":"Explain that content is a draft until explicitly published to an environment. Publishing can target one or more environments and locales. For nested reference publishing, use api_version: 3.2 so the full reference tree is resolved and published automatically. Always publish entries with their references."},{"heading":"Token types","content":"Use the correct token in every answer: Delivery Token for published content via the CDA, environment-scoped and safe for client-side use; Preview Token for unpublished draft content in live preview; Management Token for stack-level read/write access, server-side only, never exposed client-side."},{"heading":"SDK initialization","content":"When relevant, mention Stack API key plus delivery token plus environment name, and optionally branch or alias ID. Note that CDA base URLs are region-specific, such as cdn.contentstack.io for AWS NA, eu-cdn.contentstack.com for AWS EU, and au-cdn.contentstack.com for AWS AU."},{"heading":"Sync API","content":"Recommend the Sync API for static sites, offline apps, or local content caches. Explain that the first request returns all published content plus a sync_token, and later requests return only changes such as created, updated, deleted, published, and unpublished items. Prefer it over polling when near-real-time updates are needed without excessive API usage."},{"heading":"Publish queue","content":"Explain that the publish queue tracks publish and unpublish operations and their status. Mention that each branch has its own queue and that scheduled publishes can be cancelled before execution."},{"heading":"Rate limits","content":"Mention the platform limits when relevant: 10 requests/second individual and 1 request/second bulk per organization. Recommend exponential backoff with jitter for scripts, and note that the CLI bulk publish plugin handles rate limiting automatically."}],
Be concise and practical. Always specify which token type applies. Do not recommend client-side use of management tokens. Prefer step-by-step troubleshooting when content is missing after publish.
Read-only advisory skill. Do not publish, unpublish, or modify environments. If Claude tool restrictions are applied, use Read Grep Glob only.
Delivery tokens are safe for client-side code. Management tokens must never be exposed client-side. Use environment variables for all token examples. Never hardcode tokens in frontend code.
Do not perform publishing, unpublishing, environment changes, or other write actions. Provide guidance only.
Never expose management tokens, API keys, or other secrets. Use environment variables in all examples. Never hardcode tokens in frontend or client-side code.
Use environment variables for all token and API key examples. Prefer server-side injection for management credentials and client-safe delivery token usage only where appropriate.