Skip to content

Agentic Integration

cuvis-ai-agentic-skills lets assistants and automation tools interact with cuvis.AI through structured skills.

Use it to:

  • design hyperspectral pipelines from a plain-English request,
  • inspect, visualise, and execute pipelines,
  • run statistical or gradient training runs,
  • author and package custom nodes and plugins,
  • convert pipelines for cuvis.next deployment.

The skills run inside Claude Code and are distributed as a single plugin bundle from cubert-hyperspectral/cuvis-ai-agentic-skills.

Get started

  • About Skills

    What skills are, how they hand off to one another, and how cuvis-ai-env underpins the rest.

  • Available Skills

    Catalog of the eight skills: triggers, deliverables, and cross-references.

  • Invoking Skills

    Install the plugin bundle in Claude Code and trigger skills from a prompt.

Why a separate integration layer

Pipeline authoring, training, and deployment have a lot of moving parts — node specs, port wiring, execution stages, plugin manifests, trainrun YAMLs, CLI invocations. The agentic-skills bundle turns those moving parts into intentions ("build me an X pipeline that does Y") and emits the artifacts directly: a pipeline YAML, a training config, or a deployment-ready transformation. Skills compose: pipeline authoring hands off to visualisation; visualisation hands off to execution; execution hands off to cuvis.next conversion. Every skill defers runtime setup to cuvis-ai-env.