ci: release packages - #2846
Conversation
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ee4bbe4ae7
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| { | ||
| "name": "@moonshot-ai/kimi-code-sdk", | ||
| "version": "0.16.0", | ||
| "version": "0.16.1", |
There was a problem hiding this comment.
Bump the breaking SDK contract above patch
This release marks @moonshot-ai/kimi-code-sdk as 0.16.1, but the changeloged change replaces WorkspaceTrustInfo.gatedMcpServers from plain strings with structured records. Consumers pinned with ^0.16.0 can receive this patch automatically and then break wherever they render or compare server names as strings; this should not be published as a patch-level SDK update for a public API contract change.
AGENTS.md reference: AGENTS.md:L61-L61
Useful? React with 👍 / 👎.
17abaf2 to
5c09f44
Compare
5c09f44 to
28866fb
Compare
This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.
Releases
@moonshot-ai/kimi-code@0.36.0
Minor Changes
#2830
ec84a6fThanks @liruifengv! - Add an experimental fullscreen TUI mode. Set theKIMI_CODE_TUI_FULL_SCREEN=1environment variable to enable it.#2700
c9bfe8bThanks @7Sageer! - Add a configurable model pool for spawned subagents behind thesecondary-modelexperiment (KIMI_CODE_EXPERIMENTAL_SECONDARY_MODEL=1, or the master flag): with the experiment on, the/secondary-modelcommand or the[secondary_model]section in config.toml sets a default model or a small named pool that the main agent picks from per spawn. A lone legacymodelkey in the same section keeps working as the fallback default.Patch Changes
#2830
ec84a6fThanks @liruifengv! - Render LaTeX math formulas ($…$/$$…$$) in messages as Unicode formulas.#2855
30f56a2Thanks @liruifengv! - Fix Ctrl+C being ignored during automatic retries of failed API requests.#2819
fe3cdaeThanks @7Sageer! - Fix sessions failing with a provider 400 error on every follow-up request after a turn is interrupted while the model is still thinking, on strict OpenAI-compatible providers.#2847
3b0936dThanks @sailist! - Fix plain Markdown files (such as CHANGELOG.md) in an installed plugin's root directory being misidentified as skills when the plugin relies on the root SKILL.md fallback.#2843
c212ae9Thanks @sailist! - Show project MCP launch targets in the workspace trust prompt, default to declining trust, and resolve fd and stty binaries to absolute paths so untrusted workspaces cannot plant bare-name executables before confirmation.@moonshot-ai/kimi-code-sdkcontract change:WorkspaceTrustInfo.gatedMcpServersnow carries structuredWorkspaceTrustMcpServerInforecords (name,transport, andcommand/args/cwdorurl) instead of plain strings, so SDK consumers rendering a trust prompt can show the full launch target.#2856
504e629Thanks @pvzheroes125! - Refresh active MCP connections after OAuth credentials are added or reset.@moonshot-ai/kimi-code-sdk@0.17.0
Minor Changes
c9bfe8bThanks @7Sageer! - Remove the secondary-model session APISession.applyPersistedSecondaryModel; subagent model selection is configured via[secondary_model]in config.toml instead. TheSECONDARY_DERIVED_MODEL_ALIASexport stays (the v1 engine still synthesizes the entry at runtime, so hosts keep filtering it out of model pickers), and the SDK now also exportsPRIMARY_SUBAGENT_MODEL_CHOICE, the v2 subagent model pool's reservedprimarykey.Patch Changes
#2843
c212ae9Thanks @sailist! - Show project MCP launch targets in the workspace trust prompt, default to declining trust, and resolve fd and stty binaries to absolute paths so untrusted workspaces cannot plant bare-name executables before confirmation.@moonshot-ai/kimi-code-sdkcontract change:WorkspaceTrustInfo.gatedMcpServersnow carries structuredWorkspaceTrustMcpServerInforecords (name,transport, andcommand/args/cwdorurl) instead of plain strings, so SDK consumers rendering a trust prompt can show the full launch target.kimi-code@0.6.9
Patch Changes
c9bfe8b,c212ae9]:@moonshot-ai/acp-adapter@0.3.8
Patch Changes
c9bfe8b,c212ae9]:@moonshot-ai/agent-core-v2@0.3.2
Patch Changes
43c68f5Thanks @liruifengv! - Keep session updatedAt stable across metadata management writes: rename and archive/restore no longer bump it, fork inherits the source session's recency, and agent registration is non-touching; add SessionMeta.archivedAt (set on archive, cleared on restore) and surface it as archived_at through the session index and the v1/v2 session routes.@moonshot-ai/pi-tui@0.84.2
Patch Changes
#2830
ec84a6fThanks @liruifengv! - Add fullscreen groundwork to the alternate-screen renderer:ScrollView.canScroll,TuiAltScreen.getLayoutRoot(), and viewport navigation keys (PageUp/PageDown/Home/End and friends) now fall through to the focused component when the primary scroll view has nothing to scroll. Terminal focus in/out reports are no longer consumed by the viewport input handler, so app-level listeners (focus-aware notifications, clipboard-image hints) keep working in fullscreen.#2830
ec84a6fThanks @liruifengv! - Re-baseline the fork on upstream pi-tui v0.84.1 plus upstream main up to40a3d85(2026-08-11), which adds the fullscreen transcript search (ctrl+shift+f), single-line scroll actions, the ~9-18x alternate-screen render-churn reduction, and an SSH-aware escape-timeout default. The upstream renderer is now split into main-screen and alternate-screen implementations (theTUIclass is now an interface implemented byTuiMainScreenandTuiAltScreen), and the Markdown component gained opt-out LaTeX math rendering. All local patches are retained: narrow-terminal hardening, processed-line render caching, editor history hooks, the paste-burst fallback, and multi-root@completion.Editor.setTextaccepts apreservePasteRegistryoption so subclasses can replace text without orphaning live paste markers (upstream resets the registry on everysetText).