Record @figure V3 design deferred to post-custom-inlines
V3 (inline-AST-driven caption with native {image:…} detection and
future {linked-image …} custom inline) waits on the odoc custom
inline extension point, since odoc's non_link_inline_element excludes
media and the clean syntax requires {linked-image …}. V1 remains the
working implementation in the meantime.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>