setupv1.9.17

A platform team wants to run automated quality checks on generated skills or prompts in their CI pipeline without sending data to external APIs. They use this setup skill to provision an ONNX runtime environment via uv, enabling fully local inference. The daemon starts on the next session, so regression tests can evaluate output quality privately.

A healthcare analytics company needs to evaluate model outputs but cannot send patient data to cloud LLM providers. They provision a local ONNX inference environment so quality scoring stays on their infrastructure. This setup skill bootstraps the runtime with a single reproducible command.

An internal platform team maintains a private marketplace of agent skills and needs to score submissions for quality. Using this setup skill, they provision an isolated inference daemon with onnxruntime in a virtual environment. Evaluators can then run skill quality benchmarks without external network dependencies.

An IoT company ships edge devices that must rank or score model outputs locally. They use this skill to create a portable Python 3.11 environment with ONNX runtime, then bundle it into their deployment image. Because provisioning uses uv, builds remain fast and reproducible across device variants.

A seed-stage AI startup wants to test different ONNX-quantized models for output quality before committing to cloud inference costs. They use this setup skill to spin up a local runtime in minutes. The daemon starts automatically on the next session, letting founders iterate on evaluations without managing dependencies manually.

The setup skill is distributed as part of the open-source claude-night-market oracle plugin under a permissive license. It builds community adoption and drives contributions back to the project. The full experience with agents, hooks, and commands is available through the Claude Code plugin installation.

The open-source setup skill is offered free to attract developers, while a hosted version of the ONNX inference daemon provides managed scaling, monitoring, and team collaboration features. Users can start locally for free and upgrade when they need shared evaluation infrastructure.

Companies in regulated industries purchase a commercial license for the oracle inference daemon with enterprise support, audit logging, and compliance certifications. The free setup skill remains the entry point, with enterprise features gated behind a separate package or license key.