mirror of
https://github.com/MoonshotAI/kimi-code.git
synced 2026-07-10 01:39:25 +00:00
docs(changeset): simplify print mode stream-json note
This commit is contained in:
parent
ebcaa405d6
commit
98f4cfd83d
1 changed files with 1 additions and 1 deletions
|
|
@ -2,4 +2,4 @@
|
|||
"@moonshot-ai/kimi-code": patch
|
||||
---
|
||||
|
||||
Expand `kimi -p` print mode to cover (and slightly exceed) kimi-cli's non-interactive `stream-json` capabilities: emit model thinking as its own JSONL line; add `--input-format text|stream-json` (multi-turn prompts over stdin), `--final-message-only` and the `--quiet` shorthand; surface background-task/cron events as `notification` lines; report turn failures as `{"type":"error",...}` JSON with retryable provider errors mapped to exit code 75; wait for background tasks before exit (gated by `background.keepAliveOnExit`); and emit the TUI activity layer (tool progress, subagent lifecycle, warnings, skill/MCP/compaction/goal/agent-status/tool-list updates) on stdout so the stream-json output stays both complete and entirely JSON.
|
||||
Expand `kimi -p` print mode: add `--input-format text|stream-json` (multi-turn prompts over stdin), `--final-message-only`, and the `--quiet` shorthand. In `stream-json` mode the output is now entirely JSON and covers the full activity stream — model thinking, tool progress, notifications, subagent and lifecycle events, and turn errors (retryable provider errors map to exit code 75). Background tasks are drained before exit, gated by `background.keepAliveOnExit`.
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue