Inline, model-invocable skill to inspect and safely edit kimi-code's own config.toml and tui.toml (copy to a candidate, Edit, validate with kimi doctor, timestamped backup, then overwrite). Also documents tui.toml on the config files page (en + zh).
* feat(cli): add doctor command for config validation
* fix(cli): format doctor validation errors
* fix(cli): validate doctor config through SDK RPC
* chore(changeset): simplify doctor release note
This commit scaffolds the @moonshot-ai/acp-adapter package and introduces
the full ACP (Agent Communication Protocol) server implementation for
Kimi Code CLI, including:
- Scaffold @moonshot-ai/acp-adapter workspace package with build skeleton
- `kimi acp` CLI subcommand and stdout-safe logging
- ACP version negotiation and AgentSideConnection wrapper
- Auth gate for session creation
- Session lifecycle: new, list, load with history replay
- Prompt content conversion (text, image, embedded resources, resource links)
- Assistant streaming with thinking support and end-turn handling
- Tool call streaming (started, delta, progress) with result conversion (text / diff)
- Approval handling with diff/text display blocks mapped to ACP options
- Kaos read/write interface (AcpKaos) for unsaved buffer access
- Session mode (yolo/auto) and model management
- Config options builder with thinking toggle
- MCP server forwarding from ACP to harness
- Agent plan updates and available commands updates
- AskUserQuestion bridged to session/request_permission
- Plan review options surfaced through requestPermission
- Error mapping, ext_method stubs, and graceful shutdown
- IDE integration guide (Zed + JetBrains)
- End-to-end tests against ACP TS SDK client
* feat(agent-core): add dead-end detection for repeated tool calls
- inject stronger dead-end reminder at streak 10 and force-stop turn at streak 15
- emit `tool_call_repeat` telemetry events for streak >= 2
- pass telemetry client from TurnFlow to ToolCallDeduplicator
- add comprehensive tests for dead-end stop and telemetry behavior
* fix(agent-core): tune tool-dedup thresholds and preserve isError state
- Lower repeat-reminder thresholds to 3, 5, 8 and force-stop to 12
- Force-stop no longer overrides successful tool results to isError
- Move stopTurn field into ExecutableToolSuccessResult for explicit loop-control without requiring isError
- Add `applyCustomRegistryEntries` helper to apply all entries in-memory before a single write
- Update CLI Add Platform flow to use it, replacing interleaved in-memory mutations with removeProvider RPCs
- Add regression tests for repeated imports and provider field refreshing
* fix(glob): normalize pattern before brace expansion
* fix(glob): preserve backslash-escaped metacharacters and normalize braces correctly
- expand braces before normalizing to prevent `..` from collapsing across brace groups
- skip normalization for patterns containing backslash-escaped glob metacharacters
- add test coverage for escaped braces and nested `..` inside alternatives
* feat(agent-core): propagate app version into agent record metadata
- add appVersion option to AgentOptions, SessionOptions, and KimiCoreOptions
- forward appVersion through Agent, Session, and SDKRpcClient
- include app_version in agent record metadata events
Some Git for Windows installations have bash.exe only at
<root>\usr\bin\bash.exe while <root>\bin\bash.exe does not exist.
Add usr\bin\bash.exe candidates alongside each bin\bash.exe entry in
both the hardcoded fallback list and the git.exe inference logic.
Also add @moonshot-ai/kimi-code-sdk to the changeset since the SDK
bundles agent-core code and calls detectEnvironmentFromNode() on
session creation, so Windows SDK users also need this fix.
* feat(glob): add brace expansion and fix backslash escaping
- add brace expansion support for glob patterns with up to 64 sub-pattern cap
- remove up-front rejection of pure-wildcard and **/ prefix patterns; rely on 100-match cap\n- fix backslash escape handling in globPatternToRegex and inside character classes
- include matched count in truncation messages\n- show glob pattern, path, and include_dirs in TUI tool-call headers
* fix(kaos): use charAt to avoid TS strict undefined index error
* feat(tui): add /provider command, custom registry import, and tabbed model selector
- add "/provider" slash command for managing AI providers with CRUD UI
- add custom registry import via api.json URL and Bearer token
- introduce tabbed model selector grouped by provider
- add fetchCustomRegistry and applyCustomRegistryProvider in oauth package
- replace deprecated "/connect" command with unified "/provider" flow
- update provider and slash-command documentation
* feat(tui): background provider model refresh at startup
- add `refreshAllProviderModels` utility supporting managed OAuth, open platforms and custom registries
- wire background refresh into `KimiTUI` startup via `AuthFlowController`
- add `providerSwitchHint` option to `ModelSelector` and enable it in `TabbedModelSelector`
- update `TabbedModelSelector` hint wording from "switch" to "provider"
* feat(cron): render scheduled reminders in TUI and expose fired events
- add CronMessageComponent for distinct cron transcript entries in TUI
- emit cron.fired events from agent-core and expose via node-sdk
- report cron fire times with local ISO timestamps including timezone offsets
- render cron_job and cron_missed origins during session replay instead of skipping
- handle warning events in CLI and session event handler
* docs(cron): clarify that users must ask the model to manage cron tasks
- cron-create.md: instruct model to proactively tell users how to cancel/modify reminders
- cron-list.md: add guideline that users cannot directly manage cron tasks\n- cron-delete.md: emphasize users must ask model to cancel reminders
When the user interrupts running tools or parallel subagents, the tool_result
fed back to the model was a neutral `Tool "X" was aborted` or a weak "stopped by
the user", so the model treated it as a system fault and speculated about
capacity/concurrency limits instead of recognizing a deliberate stop.
Carry a UserCancellationError as the AbortSignal reason from the cancel sites
(Turn.cancel/abortTurn, SessionSubagentHost.cancelAll) through to the message
sites (tool-call settle paths and the AgentTool catches), which now emit an
explicit "deliberate user action, not a system error/timeout/capacity limit"
message. Aborts propagated from another signal (e.g. a subagent's deadline via
waitForCurrentTurn) carry their original reason, so a timeout is not mislabeled
as a user interruption. The telemetry outcome classifier matches the new
"manually interrupted" phrase to keep counting these as cancelled.
* feat: force adaptive thinking via KIMI_MODEL_ADAPTIVE_THINKING
Add the KIMI_MODEL_ADAPTIVE_THINKING env var and a matching
adaptive_thinking model-alias field that force adaptive thinking
(thinking: { type: 'adaptive' }) on or off, overriding the Anthropic
model-name version inference.
This lets custom-named staff endpoints that back an adaptive-capable
model opt in even when the model name does not encode a parseable Claude
version (which would otherwise fall back to budget-based thinking).
The kosong AnthropicChatProvider resolves the adaptive decision once in
withThinking() as `_adaptiveThinking ?? supportsAdaptiveThinking(model)`
and threads it through clampEffort/supportsEffortParam so forced adaptive
also unlocks the max effort level.
* fix: make adaptive_thinking imply the thinking capability
A model alias that sets adaptive_thinking=true (or env
KIMI_MODEL_ADAPTIVE_THINKING=true) without listing a thinking
capability previously resolved to thinking=false for custom-named
endpoints not in the capability catalog. The model picker then
classified the alias as "thinking unsupported" and forced thinking
off (persisting default_thinking=false).
Treat a forced adaptive flag as advertising the thinking capability
in resolveModelCapabilities, so the config.toml one-field opt-in
agrees with the KIMI_MODEL_* env path (which already defaults
capabilities to include thinking).
* fix: honor adaptiveThinking in the model picker, not the capability resolver
The earlier resolveModelCapabilities change had no observable effect:
modelCapabilities.thinking is consumed only for completion-budget and
media gating, never to decide whether thinking is dispatched or whether
the model picker offers a thinking toggle.
The mechanism that actually forced thinking off for an adaptive_thinking
alias is the TUI model picker, which reads ModelAlias.capabilities
directly (availableModels comes straight from config.models). Revert the
resolver change and instead make thinkingAvailability() treat
adaptiveThinking=true as a thinking toggle, so a custom-named config.toml
alias configured with only `adaptive_thinking = true` is no longer
classified "unsupported" (which forced thinking off and persisted
default_thinking=false on select). The KIMI_MODEL_* env path was already
covered by DEFAULT_CAPABILITIES.
