Drop stored pending field; incomplete turns are just history ending in
UserChatMessage. Compact seeds use an assistant summary so they are not
mistaken for retryable orphans.
Keep incomplete user messages in memory/DB; roll back only
assistant/tool on failure. /retry and resume of orphan-user sessions
re-run without duplicating the user message.
Tool-loop rounds re-send history, so summed prompt tokens are billing
not context size. Track last_request_usage and label turn/session as
billed; clarify /status and end-of-turn lines.
Max-rounds and destructive-tool prompts use asyncio.to_thread with
cancel paused on the main thread so multi-tool turns no longer look
like user cancel. Hooks may be sync or async.
Do not await async generators for raw SSE; close stream on cancel;
merge tool calls without relying only on finish_reason; quieter tool
output; /status slash command.
NCBM