Structured grounding

Method context that AI can use

Give assistants small, identifiable, connected units of method knowledge—not a pile of pages to guess from. Generated manifests, prompt packs, and graph relationships create an inspectable grounding layer.

Browse method data

For agents and assistants

Bound the context before generating the answer

Stable IDs connect human-readable content to machine-readable relationships. Clients can select the relevant role, goal, station, and resources before asking a model to assist.

Prompt packs

Localized prompt packs package reusable method instructions so integrations can begin from maintained guidance rather than improvised system prompts.

Method manifest

The MCP-oriented manifest advertises method data surfaces to tools and agents, making available context discoverable.

Graph-grounded retrieval

Resolve adjacent stations, participating roles, and supporting resources through explicit edges instead of similarity alone.

Human-verifiable output

Every unit retains a stable route and identity, so generated recommendations can point people back to the source method page.

Put it to work

Build on the published artifacts

Review the downloadable prompt packs, manifest, graph, and catalog before connecting the method to your assistant.

Browse method data