What problem does it solve? Deploying and operating apps on the Prisma Platform involves concepts that change quickly and are easy to get wrong: branches as preview environments, service versions, environment-variable classes, and secrets shown only once. This Skill gives an AI agent the authoritative, version-matched mental model of the platform so it deploys, promotes, and troubleshoots correctly instead of relying on outdated training data. ## Core Features & Use Cases - Resource model guidance: Explains workspaces, projects, branches, services, versions, Prisma Postgres databases, and S3-compatible buckets, and how the Composer module declaration drives convergence on deploy. - Deployment workflows: Covers GitHub-integration deploys, CLI deploys with prisma deploy --stage, promotion, zero-downtime rollback, custom domains, and the local prisma dev stack. - Failure-mode reference: Maps ten common failures (lost secrets, reclaimed previews, sleeping services killing background work, missing Temporal global, invalid stage names) to their fixes. - Use Case: You push a feature branch and need a preview environment with its own database. The Skill explains that the branch is the environment, that it gets a fresh empty database wired as DATABASE_URL, and that deleting the git branch tears everything down. ## Quick Start Ask the agent to deploy your Composer module to a preview stage and explain how its database and environment variables are resolved.
