Surface active ghost text suggestion in conversation context

When a user accepts or dismisses a ghost text (grey prompt) suggestion by pressing Tab or typing over it, the suggestion content is not visible in the Claude Code conversation history. If the user's response to Claude is implicitly a reaction to the ghost text (e.g. "No", "Yes", "do that"), Claude has no context for what is being responded to and must ask for clarification, adding an unnecessary round-trip.

Request: Include the active ghost text suggestion in the conversation context visible to Claude — either surfaced inline in the chat history, or passed as metadata alongside the user's next message — so that responses to suggestions are interpretable without clarification.

Example scenario:

1. Claude updates a Jira ticket with a caveat
2. Ghost text suggests: "Should I add the same caveat to all child tasks?"
3. User types "No"
4. Claude has no visibility of the suggestion and cannot interpret the response

Expected behaviour: Claude understands "No" as declining the suggested follow-on action without needing to ask "No to what?"