This repository publishes the CLARE Prompting and AI Glossary site. Visitors can browse prompt frameworks, AI terms, and the classic guided tools. The data catalog lists the classic JavaScript datasets.
GitHub Pages serves this repository's main branch from its root. index.html loads the current resource library from assets/catalog.json; the classic pages under glossary/ load their own JavaScript data files. Keep both routes working. The versioned source and build process now live in library/, with owner confirmation, exact input hashes and verified browser/SQLite projections. The deployed catalog remains a preserved snapshot; new dataset releases do not automatically replace it.
There is no root package.json or public build:data script. For a local check, serve this directory over HTTP, open the root page, and confirm that the catalog loads. Open the data catalog and a classic glossary page as separate route checks. See AGENTS.md for edit boundaries.
CLARE and the AI glossary will share the existing Activity Library Site at https://activity-atlas.jkylehobson.chatgpt.site/ai/. This repository remains the authoritative public source and corpus snapshot. Build the public projection with python3 scripts/build-site.py; the resource Site owner imports that projection under dist/ai/, preserves sibling libraries, and publishes the reviewed combined revision through the Sites helpers. This repository must not be independently deployed over the resource Site. Downloads remain on GitHub. Visitor browser data is separate from source and needs an explicit migration path before origin forwarding.
AI Glossary Library contains all 1,740 normalized records, their authoring inputs, reproducible projections and complete read-only MCP runtime. See its ownership, publication scope and reuse notice. Source releases use library/v<version> tags in this repository. The package is excluded from the static Site build and does not change existing browser or adapter data.