What problem does it solve? Developers building apps on inference.sh often misconfigure project scaffolding, resource settings, or output metadata, leading to failed deployments and silently dropped usage tracking. This Skill provides the correct CLI-driven workflow and platform-specific rules to avoid those pitfalls. ## Core Features & Use Cases - CLI-Guided Scaffolding: Enforces belt app init for creating apps with correct inf.yml, inference.py/inference.js, and dependency files. - Full Development Workflow: Covers the mandatory cycle of scaffold, implement, local test, deploy, and cloud verification with output_meta checks. - Resource & Secrets Configuration: Guides GPU/VRAM/RAM settings, secrets, OAuth integrations, and usage tracking metadata for both Python and Node.js. - Use Case: You want to wrap an external image-generation API as a CPU-only inference.sh app. The Skill provides the API-wrapper template, correct BaseAppOutput usage for output_meta, and the deploy-and-verify command sequence. ## Quick Start Ask the assistant to scaffold a new inference.sh app called my-app, implement its run function, test it locally, and deploy it with the belt CLI.
