What problem does it solve? Building Griptape Nodes that wrap an open-source ML model requires repetitive boilerplate: parameter setup, HuggingFace model selection, async inference, artifact handling, manifest registration, and lint compliance. This Skill automates that entire implementation workflow from a single spec file. ## Core Features & Use Cases - Spec-Driven Node Generation: Reads a spec file and creates node Python files using the SuccessFailureNode + AsyncResult pattern with deferred submodule imports and class-level model caching. - Built-in Parameter Components: Enforces correct use of HuggingFaceRepoParameter, SeedParameter, and ProjectFileParameter instead of raw parameters, including runtime checkpoint filename resolution. - Artifact & Media Handling: Reads and writes image, audio, and video artifacts via URL artifacts and File-based macro path resolution to avoid WebSocket saturation. - Manifest Registration & Validation: Registers each node in griptape-nodes-library.json, runs make check/make fix for lint compliance, and writes a complete library README. - Use Case: Given a spec describing an audio generation model with HuggingFace repo IDs, generate the full node file, register it in the manifest, pass pyright/ruff checks, and produce the library README in one run. ## Quick Start Run the skill with the path to your node spec file, for example: implement the nodes described in specs/my-model-spec.md.
