What probably matters to youSection of the release
What
Hooks are your own scripts that Claude Code runs at set points. A hook that asks the model a question through $.model.complete can now pass promptBlocks and systemBlocks instead of a single plain text prompt and system message. Each block is { text } and can also include cache: true.
Blocks marked cache: true are sent as cacheable blocks. Caching lets a repeated part of a prompt be reused instead of processed again.
If caching is disabled, Claude Code drops those cache markers.
Badly shaped input is rejected with specific error messages.
Why
Hook authors who send the same long text again and again can use prompt caching.
One source agreesOne thing we can check says the same as this entry.
Anthropic's release notes agreeAdded prompt caching to $.model.complete for mods: prompt and system take blocks of text, and cache: true on a block caches the request up…
