The work of defining a JSON-LD context and RDF vocabulary for GeoJSON is done. Documentation on version 1.0 of the context and context, including links to the JSON-LD and RDF docs, is published at http://geojson.org/geojson-ld/.
This work has no impact on most GeoJSON applications. For JSON-LD applications it provides a default context, allowing ordinary GeoJSON to be parsed (for the most part; see the outstanding issues noted at the end of the documentation) as though it were JSON-LD in a standard way, and helps make GeoJSON elements useful in other JSON-LD contexts.
Thanks to everyone who participated in the discussions on GitHub: Chaz6, ManoMarks, Roselin, adoyle, ajturner, akuckartz, azaroth42, calvinmetcalf, cappelaere, danbri, dinizime, dlongley, dr-shorthair, dret, ekansa, elf-pavlik, emeeks, erictheise, fils, gkellogg, hobu, jasnell, jyutzler, kgeographer, kgjenkins, ktk, lanthaler, letmaik, mitar, mpdaly, msporny, p3dr0, peterisb, pietercolpaert, retog, rtroncy, rybesh, sdrees, sfsheath, sgillies, thijsbrentjens, tschaub, and tstone. Because I drastically reduced the scope of the work toward the end, some of these contributions aren't reflected in the final products.
A vocabulary for describing the temporal extent of event-like features is one of the things that I cut from GeoJSON-LD. It lives on as a proposed extension for RFC 7946 GeoJSON, with no JSON-LD requirements, at https://sgillies.github.io/geojson-events/.
Work on resolving the mismatch between GeoJSON's nested coordinates array and JSON-LD could be part of the JSON-LD 1.1 discussion. Make sure to subscribe to email@example.com and https://github.com/json-ld/json-ld.org if you're interested in this issue.