* feat: polish vis
* feat: add 'kimi vis' command for session visualization
* fix(vis): drop metadata app_version/resumed removed upstream
#786 stopped recording resume version metadata, so those fields no
longer exist on the metadata wire record. The vis-into-typecheck wiring
caught the stale field reads after merging main; drop them from the
metadata headline.
* fix(vis): drop unnecessary return-await in startVisServer
oxlint typescript-eslint(return-await) flags returning an awaited
promise outside try/catch; return the promise directly.
* fix(vis): green CI — tolerate unbuilt embedded asset + bump nix pnpmDeps hash
- handleVis: wrap the embedded-SPA dynamic import in try/catch. The value
module is generated at build time (prebuild); in contexts without a build
(tests run pnpm test, not build) only the .d.ts type stub exists, so the
runtime import throws. Tolerate it and fall back to filesystem serving.
- flake.nix: update the fetchPnpmDeps hash after adding the vis-web /
vis-server / vite-plugin-singlefile dependencies.
* refactor(vis): drop redundant alwaysBundle in tsdown config
#775's single-entry build (codeSplitting: false) already bundles
everything not declared in dependencies/peerDependencies. hono /
@hono/node-server (transitive via vis-server) and @moonshot-ai/vis-server
(a devDependency) are all undeclared there, so they bundle by default —
the explicit alwaysBundle was redundant. Verified the emitted main.mjs is
still fully self-contained and 'kimi vis' serves.
* fix(vis): address review — context-token resets, IPv6 url, marker-safe indexing
- contextTokens now mirrors agent-core on lifecycle records: 0 on
context.clear, tokensAfter on context.apply_compaction (was only
updated from step.end.usage, leaving a stale live fill after a
clear/compaction).
- start.ts brackets IPv6 hosts in the returned url (http://[::1]:port/);
hostForUrl moved to config.ts and shared with the startup banner.
- compaction slice + micro-compaction blanking now index over real
history entries only, so synthetic undo/clear UI markers no longer
offset agent-core's compactedCount / cutoff.
* chore: add changeset for kimi vis command
* fix(vis): cross-platform single-file build + history-count micro clamp
- build-vis-asset.mjs sets VIS_SINGLEFILE via the spawn env and runs
'vite build' directly (cross-platform), instead of the POSIX-inline-env
'build:single' script that broke on Windows cmd; removed the now-unused
build:single script. Fixes the win32 build path (the asset generator
runs in the kimi-code prebuild + native bundle).
- context.undo now clamps the micro-compaction cutoff by history-entry
count (excluding synthetic undo/clear markers) instead of messages.length,
mirroring agent-core undo() -> microCompaction.reset(_history.length); a
surviving marker no longer leaves the cutoff one too high and wrongly
blanks a later-appended tool result.
* fix(vis): run the single-file build through a shell for Windows pnpm
The win32 native binary is built on Windows runners
(.github/workflows/_native-build.yml), which run this generator. pnpm's
launcher there is pnpm.cmd, which a bare argv exec can't resolve without
a shell. Use execSync with a single command string so the platform shell
(cmd on Windows) resolves the shim; a command string (not an args array)
avoids the args+shell deprecation. Args are static.
* fix(vis): show model-facing tool result content in the context view
agent-core normalizes tool results via toolResultOutputForModel before
they enter history (error -> '<system>ERROR: ...' prefix, empty ->
'<system>Tool output is empty.' sentinel). The projector was using the
raw ev.result.output, so the Context tab's model view showed content the
model never saw for failed/empty tool calls. Replicate that normalization
(the upstream helper is module-private) so the projected tool message
matches what the model received.
* fix(kosong): isolate anthropic auth environment
* fix(kosong): close remaining anthropic env fallbacks
Pass explicit nulls into the Anthropic SDK for unused auth/base URL overrides, keep adapter-owned auth headers authoritative, and add regression coverage for Anthropic shell env leakage.
* fix(kosong): block anthropic custom header env leakage
* docs(kosong): explain anthropic env-isolation intent + migration
Spell out that the SDK is used as a transport to arbitrary endpoints, so disabling its shell-env auto-discovery is the fix: the authToken/baseURL/header nulls are load-bearing, not redundant. Also document the behavior change (shell ANTHROPIC_* no longer read; use provider config) in the changeset. No logic change.
* docs(changelog): sync 0.15.0 from apps/kimi-code/CHANGELOG.md
* docs(changelog): improve Chinese translation for 0.15.0 tool-call status entry
* docs(changelog): polish Chinese wording for 0.15.0 entries
* feat: add all-sessions picker with name search
* fix: satisfy clipboard lint errors
* test: make clipboard fallback assertion portable
* fix: harden session picker interactions
* fix: harden session picker startup flow
Invalidate pending session picker scope changes after dismissal and exit startup picker after showing a cross-directory resume command.
* fix: show empty session scope hint
Render the all-sessions toggle hint in the sessions picker empty state so users can discover Ctrl+A when the current directory has no sessions.
* fix(kimi-code): bundle npm package into single entry
Bundle runtime dependencies into the CLI npm entry file, prepare an npm publish directory without bundled dependencies, and keep clipboard support optional.
* fix(kimi-code): sync npm publish directory lockfile
Update pnpm-lock.yaml for apps/kimi-code publishConfig.directory so CI can install with --frozen-lockfile.
* fix(kimi-code): disable cli declaration output
Stop generating the TypeScript declaration sidecar so the npm package contains only the main.mjs CLI entry artifact.
* fix(kimi-code): move bundled deps to devDependencies
Remove the npm publish directory pre-processing path, keep runtime dependencies as dev-only build inputs for the single-file bundle, and add a manifest regression test.
* fix(kimi-code): make koffi optional
Keep koffi available as an optional native dependency without listing bundled runtime packages as published dependencies.
* fix(kimi-code): use explicit bundle allowlist
Avoid treating every devDependency as a runtime bundle input and list the bundled third-party dependencies explicitly.
* fix(kimi-code): remove manifest regression test
Keep the packaging change focused on the manifest and bundler configuration without an extra package-json regression test.
* fix(kimi-code): only externalize optional deps
Keep tsdown's never-bundle list limited to published optional native dependencies.
* Update tsdown.config.ts
* fix(kimi-code): simplify changeset
Keep the changeset summary focused on optimizing the npm packaging system.
* feat(agent-core): collapse hidden dirs in prompt
Keep hidden directory entries visible in the cwd snapshot while omitting their contents to reduce prompt noise, and document the tools models should use to inspect hidden paths.
* feat(agent-core): clarify prompt AGENTS context
Add a marker when AGENTS.md content is truncated by the prompt budget and move dynamic system prompt context after the static guidance.
* chore: add prompt refinement changesets
* fix: prompt brief tool-call status
Ask the model to emit one short same-language status sentence before non-trivial tool calls while keeping detailed reasoning out of the visible transcript. Updates snapshots for the longer default prompt.
* feat(agent-core): collapse hidden dirs in prompt
Keep hidden directory entries visible in the cwd snapshot while omitting their contents to reduce prompt noise, and document the tools models should use to inspect hidden paths.
* feat(agent-core): clarify prompt AGENTS context
Add a marker when AGENTS.md content is truncated by the prompt budget and move dynamic system prompt context after the static guidance.
* chore: add prompt refinement changesets
Replace the per-provider getCapability instance method, and the throwaway
provider instantiation used only for capability probing, with a static
getModelCapability(wire, model) entry point in kosong. The same
capability-registry tables are consulted, so resolution stays behaviorally
identical; it no longer constructs a temporary provider or forges an API key.
Also drop the unused getContextSizeLimit interface method.
* feat(update): roll out automatic updates in staged batches via CDN manifest
* chore: remove changeset
* refactor(update): single-source the CDN latest file names
* refactor(update): reuse the CDN latest URL constants in update checks
* refactor(update): drop the test-only CDN base override
* fix: preserve first launch attribution
* fix: use refreshed rollout manifest for telemetry
* fix: abort hung cdn update checks
* fix: preserve update cache with bad manifest
* chore: remove review report from branch
Resuming a session that was already in plan mode with --plan crashed with
"Already in plan mode": the resume path called setPlanMode(true)
unconditionally while session replay had already restored the active plan
state. Check the session status first and only enable plan mode when it is
not active yet, in both the resume startup path and the startup session
picker.
The /sessions picker shared the same onSelect callback, so startup flags
were also re-applied on every mid-session switch, overriding the picked
session's own persisted modes. Gate the flag application behind an
applyStartupModes option that only the startup picker enables, and surface
post-switch setup errors instead of leaving them as unhandled rejections.
* feat(providers): sync custom registry providers on startup refresh
- group registry providers by URL and retry available API keys\n- automatically add new providers and remove disappeared ones\n- coalesce duplicate source URLs to avoid false config-change reports\n- clear defaultThinking when default model is removed\n- update docs and add tests for registry sync scenarios
* fix(tui): only show provider refresh status for added models
Skip removed / metadata-only provider updates when reporting model list changes.\n\n add: test to enforce the behavior.
iTerm2 interprets any OSC 9 payload as a desktop notification, so the
ConEmu-style 9;4 progress sequence (re-sent every second by the progress
keepalive) flooded users with notifications. Only emit progress on
terminals known to implement OSC 9;4: Windows Terminal, ConEmu, Ghostty,
and WezTerm.
* feat(config): tolerate invalid config.toml sections instead of failing startup
Schema errors now drop only the offending sections (single entries for
providers/models) with a warning, so a typo no longer prevents startup or
drops the login state. TOML syntax errors still fail fast with the parse
location. Mid-run reloads keep the last good config when the file breaks.
Warnings surface via the new getConfigDiagnostics API: as a startup notice
in the TUI, on stderr in print mode, and in the status bar after /new.
Write paths stay strict so a broken file is never silently rewritten, and
now fail with a short actionable message instead of raw validation JSON;
the /provider TUI flow and the kimi provider CLI report these errors
instead of crashing on an unhandled rejection.
* fix(config): keep entry-keyed sections when one entry has multiple issues
A providers/models entry with several validation issues was deleted by the
first issue, and the remaining issues from the same safeParse pass then
escalated to deleting the entire section — one badly-typed custom provider
could drop every provider, including the managed OAuth login. Issues on
entry-keyed sections now only ever target the entry itself; once it is
gone, later issues are no-ops.
* feat(skill): qualify sub-skill names with parent prefix
- qualify sub-skill names with parent prefix and set isSubSkill metadata\n- hide sub-skills from the model skill listing
- expose sub-skills as dotted slash commands in TUI
- update slash command docs for English and Chinese
- align built-in sub-skill local names with their directories
