* feat(kosong): classify HTTP 413 request-body-too-large as a dedicated error type
* feat(agent-core): lower default image downscale cap to 2000px and make it configurable
* feat(agent-core): strip media to text markers and retry when the compaction request is too large
* feat(agent-core): cap model-initiated image reads with a configurable byte budget
* feat(agent-core): resend with degraded media when the provider rejects the request body as too large
* test(agent-core): add explicit timeouts to encode-heavy image budget tests
* feat: add WebP decoding support with wasm integration
- Introduced a new WebP decoding module using @jsquash/webp's wasm decoder.
- Implemented functions to decode WebP images and check for animated WebP formats.
- Updated image compression tests to include scenarios for WebP handling, including encoding and decoding.
- Enhanced error handling for API request size limits to accommodate various error messages.
- Updated pnpm lockfile to include new dependencies for WebP encoding and decoding.
* chore(changeset): consolidate this PR's entries into one
* fix(nix): update pnpmDeps hash for merged lockfile
* feat(agent-core): refuse HEIC/HEIF reads with platform-matched conversion guidance
* feat(kosong): honor explicit anthropic max output override
Add claude-opus-4-8 output ceiling and treat explicit max_output_size/KIMI_MODEL_MAX_OUTPUT_SIZE as the final Anthropic max_tokens value. Sync configuration and environment variable docs.
* feat(kosong): drop claude-opus-4-8 ceiling
Revert the newly added claude-opus-4-8 default output ceiling while keeping explicit max_output_size overrides for Anthropic.
* fix(kosong): fall back to nearest lower catalogued minor for Claude max tokens
An uncatalogued Claude minor version (e.g. claude-opus-4-8) previously
dropped straight to the family/major baseline ceiling, so Opus 4.8
resolved to 32k max output tokens instead of the 128k it supports.
Walk the minor version down to the nearest catalogued entry before
using the family baseline, so a newer minor inherits its predecessor's
documented ceiling.
* fix(kosong): catalogue Opus 4.8's documented 128k output ceiling
Known models should resolve from explicit table entries; the
nearest-lower-minor fallback now only covers minors that are not yet
catalogued. Fable 5 already has an explicit entry.
* feat(agent-core): record llm request trace in wire.jsonl
Add three observability record types so every request sent to the model
can be reconstructed from the wire log at the logical-request level:
- llm.tools_snapshot: content-addressed snapshot of the top-level tools
table as sent (post deferred-strip), written once per unique table
- llm.request: one record per outbound request (retries, strict resends,
and compaction rounds included) carrying the effective request params
and hash links to the system prompt and tools snapshot
- mcp.tools_discovered: the server's verbatim tools/list result plus the
agent's gating (allow-list, collisions), deduplicated by content hash
Observability records never feed state rebuild; replay only restores the
write-dedup cursors. The records/types.ts contract now documents the two
record classes explicitly (persisted is not the same as replayed).
Recording happens at the single Agent.generate choke point. The
LLMRequestLogFields side channel gains kind/projection/maxTokens/
droppedCount, chatWithRetry preserves caller-set fields, and compaction
tags its requests. The vis wire view renders the new record kinds.
* fix(agent-core): record the provider-clamped completion cap in the request trace
The llm.request trace recorded the client-requested budget cap, but
chat-completions providers tighten the actual wire value inside
withMaxCompletionTokens (remaining-context sizing, transport ceilings,
model-default resolution) — with the default budget the clamp is active
on nearly every non-empty-context request, so the recorded value did not
match what was sent.
Providers now expose the effective cap they computed as a readonly
maxCompletionTokens field on the clone, and the recorder reads it from
the effective provider at the Agent.generate choke point. This replaces
the side-channel recomputation, which is removed along with the
appliedCompletionBudgetCap helper.
* fix(agent-core): park pre-replay MCP discovery records and hash the collision outcome
Two wire-hygiene fixes for the mcp.tools_discovered trace:
Parking: the real Session ordering connects MCP servers concurrently with
agent construction, so ToolManager can observe a connected server before
agent.resume() has replayed the wire. Recording at that point bypassed
the restored dedup cursor (duplicating a 1-50KB record on every resume)
and appended a stray metadata record ahead of replay. AgentRecords now
exposes a one-shot opened latch — set when replay completes (after the
migration rewrite flushes) or when the first live record is logged — and
ToolManager parks discoveries until then, re-running the dedup check at
drain time. A frozen range-limited replay never opens; those agents are
transient previews.
Collision hashing: the dedup hash now covers the collision outcome, not
just the raw list and allow-list. Collisions depend on which other
servers hold a sanitized qualified name at registration time, so a
server can re-register with identical tools but a flipped outcome; that
gating change must produce a new record instead of being suppressed.
* fix(agent-core): skip the request trace for pre-flight-aborted calls
Mirror kosong generate()'s pre-flight abort check at the Agent.generate
choke point: a call whose signal is already aborted never reaches the
wire (generate throws before dispatching), so it must not leave an
llm.request/llm.tools_snapshot trace or a diagnostic log line claiming a
request was sent. Recording stays before dispatch for every call that
passes the gate, preserving the crash-safety of the trace.
* chore(agent-core): remove a leftover adaptive-thinking override hook
The adaptiveThinkingOverride option was a temporary local hook explicitly
marked for removal before commit. Nothing passes it, so resolution falls
back to the alias-level adaptiveThinking value in all cases; drop the
option and the dead indirection.
* fix(kosong): derive the exposed completion cap from generation kwargs
maxCompletionTokens was a field stored only by withMaxCompletionTokens,
so caps that reach the wire through other paths were invisible to the
request trace: with completion budgeting disabled via env, Anthropic
still sends the constructor-resolved max_tokens (required by the
Messages API), and constructor-level kwargs like OpenAILegacyOptions
maxTokens were likewise unreported.
Replace the stored field with a getter derived from each provider's
generation kwargs — the single source the request body reads — covering
constructor defaults, direct withGenerationKwargs configuration, and
budget application in one place. Kimi mirrors its request-time legacy
max_tokens alias normalization; openai-legacy reuses the same
normalizeGenerationKwargs the request path uses.
* feat(agent-core): add thinkingKeep passthrough for Kimi providers and update tests
* feat(agent-core): enable Preserved Thinking by default on the Anthropic provider
Default thinking.keep to "all" for the Anthropic provider (Claude and Kimi in Anthropic-compatible mode) while Thinking is on, via a context_management clear_thinking_20251015 edit, mirroring the Kimi default. Reuses [thinking] keep and KIMI_MODEL_THINKING_KEEP (env > config > default "all"); off-values disable it.
* feat(kosong): route Anthropic Preserved Thinking through the beta Messages API
Force the beta endpoint (client.beta.messages.create) when thinking.keep is enabled, since clear_thinking_20251015 is only honored there. Also prepend clear_thinking to any existing context-management edits (for example clear_tool_uses) instead of replacing them, keeping it first as Anthropic requires when combining edits.
* docs: clarify Anthropic beta endpoint and compaction keep behavior
Note in code comments and bilingual docs that enabling Anthropic Preserved Thinking routes requests to the beta Messages API (client.beta.messages.create), with keep=off as the escape hatch back to the standard endpoint. Correct the resolveThinkingKeep comment to reflect that compaction shares ConfigState.provider and intentionally carries the same keep.
* test(kosong): cover Anthropic beta endpoint (streaming and forced betaApi)
Add a streaming beta-endpoint capture and a test that withThinkingKeep forces the beta endpoint even when constructed with betaApi: false, pinning down the documented behavior.
* feat(agent-core): progressive tool disclosure via select_tools
Keep MCP tool schemas out of the immutable top-level tools[] and let the
model load them on demand, preserving the provider prompt cache:
- kosong: Message.tools (append-only load primitive, serialized as Kimi
messages[].tools with type:function wrapping and no content),
Tool.deferred (stripped once in generate() so loaded tools stay
executable without re-entering the top level), select_tools capability
bit (UNKNOWN/catalog default false).
- select_tools builtin: load-by-exact-name, three-branch semantics
settled per name (Loaded / Already available / Unknown), schemas read
from the live registry, injection-origin schema messages survive undo.
- ToolsDiffInjector: <tools_added>/<tools_removed> announcements at turn
boundaries and post-compaction, folded from history (undo/compaction/
resume self-heal), appended only when the loadable set changes.
- Loaded-tools ledger = history scan + defer-window pending set (cleared
on /clear); loop re-reads the executable table per step so a selected
tool dispatches on the next step of the same turn; preflight
distinguishes not-loaded from loaded-but-disconnected.
- Cross-cuts: projection strips protocol context for non-select_tools
models (lossless mid-session model switch both ways), compaction
filters it from the summarizer input and rebuilds loaded schemas
keep-all after folding, token estimation counts message.tools, request
logging reflects the post-strip wire tools.
- Three-condition gate: capability.select_tools x capability.tool_use x
tool-select experimental flag (KIMI_CODE_EXPERIMENTAL_TOOL_SELECT).
Any gate closed reproduces the inline request byte-for-byte; all
current models keep the capability off, so behavior is unchanged
until a supporting model is catalogued. The SDK catalog-to-alias
mapping forwards the capability so catalog-driven setups can enable it.
* feat(kosong): skip tool-declaration-only messages in non-Kimi providers
Message-level tool declarations (messages[].tools) are a Kimi wire
feature. The other providers' explicit field construction already keeps
the tools field off the wire, but the content-free leftover message
would be rejected (OpenAI: system message without content) or serialize
as a garbage <system></system> turn (Anthropic/Google system-to-user
wrapping). Skip such messages entirely via a shared predicate; a message
that also carries content only loses the tools field, as before.
Unreachable in kimi-code (the projection gate strips dynamic-tool
context for models without the select_tools capability before any
provider sees it) — defense-in-depth for direct kosong consumers.
* fix(agent-core): survive runtime flag flips and align tool table with post-compaction state
Two fixes from PR review:
- Register select_tools unconditionally and gate only its exposure in
loopTools. The tool-select flag can flip at runtime (config reload
calls setConfigOverrides on the live resolver) without
initializeBuiltinTools re-running; previously the disclosure shape
activated while the tool itself was unregistered, cutting the session
off from MCP entirely until a model/cwd change rebuilt the builtins.
A profile listing the name explicitly still never surfaces it in
inline mode, and execution guards the flip race defensively.
- Resolve the per-step tool table AFTER beforeStep, next to
buildMessages. beforeStep can run full compaction, which trims loaded
schemas and rewrites the ledger; a table captured before it could
still dispatch a tool whose schema the model no longer has. The
executable table and the request messages now always reflect the same
state, so a trimmed tool is rejected with select guidance instead of
executed.
* fix(agent-core): drop unused Tool import in dynamic-tools
* fix(agent-core): baseline compaction guard after post-compaction reinjection
The reinjected reminders (loadable-tools manifest, goal) are re-appended
after every compaction, but the nothing-new-since-compaction baseline was
captured before injectAfterCompaction. With a large manifest the guard
could re-trigger auto-compaction against a floor that cannot shrink.
Raise the baseline to the true post-compaction floor once reinjection
completes; the earlier capture stays as a fallback when reinjection
throws.
---------
Co-authored-by: fengchenchen <fengchenchen@moonshot.ai>
- send tool declarations, system prompt, and sampling/thinking settings in the camelCase shape the Google SDK forwards so tool calls reach the model
- thread tool-call extras (thought signatures) through the loop tool-call event into context so Gemini 3 can resume a tool turn
- update and add tests for the corrected request shape and signature round-trip
* feat(agent-core): guide the model away from repeating denied or failed tool calls
- system.md: add a diagnose-before-retrying paragraph next to the existing
permission-denial guidance, covering failed tool calls
- permission: when the user rejects an approval on the main agent, tell the
model not to re-attempt the exact same call (sub agents already had an
equivalent hint)
* fix(agent-core): close abandoned tool exchanges and dedupe duplicate tool_use ids
A turn that dies between a recorded tool.call and its paired tool.result
(e.g. a transcript write failure mid-batch) used to leave
pendingToolResultIds open forever: every later message was stranded in
deferredMessages and user input was silently swallowed.
- runOneTurn now defensively closes any dangling tool calls when a turn
ends (completed, cancelled, or failed), synthesizing an error result
that names the cause, with a warn log and a tool_exchange_abandoned
telemetry event
- the projector drops assistant tool calls whose id already appeared
earlier (first occurrence wins): a duplicate id is wire-invalid on
strict providers and not repairable by the strict resend; reported via
the existing projection-repair log and telemetry
- resume-side closePendingToolResults now logs what it closes (warn for
a mid-history gap, info for the routine trailing interruption)
* chore: add changesets for tool exchange fixes
* fix(agent-core): scope duplicate tool_use id dedup to the strict resend
Unconditional dedup regressed providers that emit per-response counter
ids (e.g. call_0 in every step) and accept their own duplicates: later
tool exchanges silently vanished from the projected history, and a
duplicate call's own recorded result was left dangling.
- the dedupe pass is now opt-in via dedupeDuplicateToolCalls and enabled
only in strictMessages, so the normal projection keeps the history the
provider produced
- the pass also drops every tool result after the first for an id, so no
dangling tool message survives; when the kept call has no result of
its own, the surviving one is reattached by the adjacency repair
- kosong now classifies the Anthropic "tool_use ids must be unique" 400
as a recoverable request-structure error so it triggers the strict
resend
* fix(agent-core): recover sessions bricked by orphan tool results
A stray `tool` message with no preceding assistant `tool_calls` permanently
bricked a session on OpenAI-compatible providers: every turn re-sent the same
malformed history and got a 400, and switching model/provider did not help.
Two independent gaps caused this:
- kosong did not recognize the OpenAI / DeepSeek / vLLM / Qwen phrasings of the
tool-exchange structural 400 (`role 'tool' must be a response to a preceding
message with 'tool_calls'` and the mirror `assistant message with 'tool_calls'
must be followed by tool messages`), so the post-400 strict-resend fallback
that drops the orphan never fired.
- The legacy-restore compaction path kept a verbatim tail
`history.slice(compactedCount)`; when the cut landed inside a tool exchange the
tail began with an orphan tool result whose assistant was summarized away. The
normal projection does not repair a leading orphan, so the malformed history
was baked in and re-sent every turn.
Recognize the additional phrasings so the strict resend un-bricks any session,
and trim leading tool results from the legacy-restore tail so the orphan is
never persisted in the first place.
* fix(agent-core): drop orphan tool results at the projection boundary
Rework the legacy-restore half of the fix based on review feedback: mutating
`_history` at restore time desyncs every consumer that models the history from
the wire records — the transcript reducer's fold length would overcount and
make MessageService skip unflushed live-tail messages.
Keep the restored history faithful to the wire records instead, and drop a
`tool` result whose call is nowhere in the history at the projection boundary,
on every request-building projection: the normal wire (`messages`), the
post-400 strict resend (`strictMessages`), and the compaction summarizer. An
orphan is wire-invalid on strict providers and useless to the model either
way, so it never reaches a provider — no longer relying on recognizing the
provider's 400 phrasing to recover. Fragment projections (e.g. token-estimating
a history slice) leave results untouched, since a matching call may
legitimately sit outside the slice.
Moonshot / Kimi (OpenAI-compatible) rejects a history whose tool message
references a tool_call_id with no matching tool_calls entry in the preceding
assistant message as `400 tool_call_id is not found`. The
TOOL_EXCHANGE_ADJACENCY_MESSAGE_PATTERNS only covered Anthropic's
tool_use/tool_result phrasing, so isRecoverableRequestStructureError returned
false, the strict-resend fallback in executeLoopStep never fired, and the
session stayed permanently stuck re-sending the same rejected history every
turn (observed in the field after a manual compaction busted the prompt cache
and forced full revalidation of a latently misordered prefix).
Add the tool_call_id-anchored pattern so the whole recovery chain — strict
projection (adjacency repair, orphan-result drop, synthetic results) plus the
one-shot resend — now also covers the default provider. Covered by classifier
unit tests and an e2e resend-and-recover case.
A raw undici `terminated` error — an SSE/HTTP response body cut mid-flight,
common on long streaming responses — fell through convertAnthropicError to a
generic base ChatProviderError, which isRetryableGenerateError treats as fatal,
so it was never retried. Route raw non-SDK errors through the shared
classifyBaseApiError heuristic (already used by the OpenAI path) so a dropped
stream is classified as a retryable APIConnectionError and retried instead of
failing the turn.
* feat: add KIMI_MODEL_THINKING_EFFORT to force a thinking effort
Send thinking effort only when the model declares it in support_efforts, and add the KIMI_MODEL_THINKING_EFFORT environment variable as an escape hatch to force a specific effort regardless of declared support.
* test: align thinking effort expectations with support_efforts gating
Update the kimi adapter e2e and compaction tests that asserted the previous pass-through behavior on models without support_efforts.
* feat: add model alias overrides
Preserve user model overrides across provider catalog refreshes and resolve effective model metadata for runtime, TUI, protocol, and ACP consumers.
* fix: apply model display name overrides
Show overridden model display names in the footer, welcome panel, status output, and model switch confirmations.
* fix: pass through kimi effort when undeclared
Keep support_efforts authoritative when declared, but pass requested Kimi thinking effort through when the model does not declare support_efforts.
* fix: honor model overrides in effort commands
Use effective model metadata for /effort choices and for always_thinking clamping when resolving thinking effort.
* fix(provider): honor base_url for google-genai and vertexai providers
The google-genai and vertexai provider types silently ignored a configured
base_url and always hit generativelanguage.googleapis.com (e.g. a Gemini-
compatible proxy URL + key could not be used). Plumb the endpoint through to
the @google/genai SDK via httpOptions.baseUrl:
- kosong: add baseUrl to GoogleGenAIOptions and inject it into the client's
httpOptions alongside the existing headers (the SDK merges headers and
overrides the base host).
- agent-core: forward provider.baseUrl in the google-genai and vertexai
branches, with GOOGLE_GEMINI_BASE_URL / GOOGLE_VERTEX_BASE_URL env
fallback. vertexai keeps deriving location from an aiplatform host.
- docs: document base_url for both providers, noting the host root only
must be given because the SDK appends the API version itself.
Covered by unit tests asserting the URL reaches the kosong config and the
SDK client's httpOptions.
* fix(provider): use the effective base_url for vertex location detection
The vertexai branch forwarded the endpoint from config `base_url` OR the
GOOGLE_VERTEX_BASE_URL env fallback, but service-account detection
(`hasVertexAIServiceEnv` / `vertexAILocation`) still derived the region from
`provider.baseUrl` only. Supplying the regional endpoint via the env fallback
(with a project but no explicit GOOGLE_CLOUD_LOCATION) therefore left location
undefined and silently downgraded Vertex ADC to API-key Gemini routing.
Resolve the effective base URL once and use it for both forwarding and location
derivation, so the env fallback behaves exactly like `base_url`. Add a
changeset for the kosong + agent-core patch release.
* feat(agent-core): rework compaction to keep only user prompts and summary
* refactor(agent-core): rewrite compaction summary as first-person handoff
Rework the full-compaction summary to read as the agent's own continuing
notes instead of a third-party report:
- compaction-instruction.md: free-form first-person continuation that
preserves exact commands, paths and outcomes, states the precise next
action, and flags claimed-but-unverified work rather than trusting it.
- compaction-summary-prefix.md: skeptical "your own working notes"
framing; drop the collaborative third-party prefix.
- system.md: add compaction-awareness guidance so the model continues
naturally from a summary and re-checks any reported "done".
- Rename the compaction helpers module to handoff.ts.
Update tests and regenerate snapshots for the new prompt text, and fill
in contextSummary in the restored-compaction replay expectations.
* fix(agent-core): count image/audio/video parts in token estimation
estimateTokensForContentPart returned 0 for image_url/audio_url/video_url,
so auto-compaction triggers, the overflow-shrink budget, the kept-user
budget, and the reported context size all went blind to media — a
media-heavy session could overflow the model window while the estimate
reported a near-empty context. Media parts now carry a fixed estimate
(MEDIA_TOKEN_ESTIMATE), and the content-part switch is exhaustive so a new
ContentPart kind must declare its estimate rather than silently count as
zero.
* feat(agent-core): re-surface active background tasks after compaction
Folding the live context to [recent user prompts, summary] drops the
messages that started background tasks and their status updates, so the
model could forget a task is still running and spawn a duplicate.
injectAfterCompaction now appends a system-reminder listing active
background tasks (with guidance to use TaskOutput/TaskList/TaskStop
instead of re-spawning). It runs only post-compaction and carries an
injection origin, so the next compaction drops and rebuilds it rather
than stacking copies; the all-user-role post-compaction shape is
preserved (no tool-pairing reintroduced).
* test(agent-core): add compaction scenario guards and risk probes
Adds compaction-scenarios.test.ts driving the real Agent/ContextMemory/
FullCompaction machinery:
- A guard test locking in that repeated compaction folds the prior summary
into the new one instead of stacking two summaries.
- Seven `it.fails` probes that executably reproduce known, currently-accepted
edge-case defects so the suite stays green while documenting each one
precisely; any of them will flip red (forcing removal of `.fails`) the day
the behavior is fixed. They cover: assistant/tool appended during an
in-flight summarizer call being dropped; unbounded shrink on empty
summaries; the fixed 20k kept-user budget overflowing a small model window;
a tool result orphaned when compaction starts mid-exchange; legacy
compaction records dropping their verbatim tail on replay; micro-compaction
clearing recent tool results in an overflow-shrunk suffix; and media being
discarded when the oldest kept user message is truncated.
* fix(agent-core): repair tool_use/tool_result adjacency in projected context
A tool call and its result can end up non-adjacent in history — a
background-task notification or flushed steer lands between them, or an
interrupted/nested step delays the result — which strict providers reject
with HTTP 400. The projector now moves each tool_use's result up to
immediately follow it (projection-time only; the stored history is
untouched), and full compaction projects its summarizer input with a
synthetic result for any still-open call so the summary request stays
well-formed. Micro-compaction only surfaced this latent ordering by busting
the prompt cache, so it now defaults off.
Includes projector adjacency regression tests, a context-level integration
test, and a compaction synthesize-missing guard; the prior "keeps an
unresolved tool exchange out of the compaction prompt" test is updated to
the now-well-formed (synthetic-result) behavior.
* fix(agent-core): preserve the verbatim tail when restoring legacy compactions
A pre-rework `context.apply_compaction` record used
`[summary, ...history.slice(compactedCount)]` semantics and kept a verbatim
recent tail, but it has no `keptUserMessageCount`. The reworked applyCompaction
re-folded such records into the all-user shape, dropping the recent
assistant/tool tail — so resuming a session compacted by an older version
silently lost its most recent context.
On restore of such a record (gated on records.restoring, no keptUserMessageCount,
and compactedCount < history length) reproduce the old shape instead. The
forward/live path is unchanged; the projector's tool-adjacency repair keeps the
restored tail well-formed, and compaction only runs at clean step boundaries so
the tail has no open exchange. The legacy-tail probe now passes as a regression
guard via the real restore path.
* fix(agent-core): align legacy compaction foldedLength with live restore
The transcript reducer re-derived foldedLength for pre-rework
context.apply_compaction records (no keptUserMessageCount) using the new
kept-user+summary rule, but ContextMemory's restore now reproduces the legacy
[summary, ...history.slice(compactedCount)] shape for those records. The two
diverged for legacy sessions, so MessageService's foldedLength-vs-live-history
comparison could mis-handle GET /messages (miss or misorder recent output).
The reducer now mirrors the live legacy fold: when compactedCount is below the
pre-compaction length it computes 1 + (length - compactedCount); otherwise it
falls back to the kept-user derivation. The MessageService transcript test's
fixture is corrected to a new-format record, matching its all-user live mock.
* fix(kosong): merge a follow-up user turn into the preceding tool_results
The Anthropic message merge keyed on isToolResultOnly(last) ===
isToolResultOnly(converted), which left a tool_result-only user turn
followed by a plain-text user turn unmerged. After tool-exchange repair
this shape (assistant tool_use -> tool_result -> injected notification)
produces two adjacent user messages, which strict Anthropic-compatible
backends reject with HTTP 400.
Switch to the asymmetric predicate isToolResultOnly(last) ||
!isToolResultOnly(converted): a tool-result-only running message absorbs
whatever user turn follows (parallel tool_results or a trailing text),
yielding a valid [tool_result, ..., text] message; a plain-text running
message still only absorbs plain text. [tool_result, text] is valid for
both native Anthropic (which concatenates anyway) and strict backends.
* test(agent-core): pin micro-compaction flag in the shrunk-suffix probe
The 'does not clear recent tool results when projecting a shrunk suffix'
probe is an it.fails that only documents a real defect while
micro-compaction is active. It inherited the ambient
KIMI_CODE_EXPERIMENTAL master switch, so its pass/fail flipped with the
runner: green locally (master switch on) but a hard failure in CI, where
the flag defaults off and MicroCompaction.compact() is a no-op that
leaves the tool result intact.
Enable KIMI_CODE_EXPERIMENTAL_MICRO_COMPACTION explicitly for this probe
so it deterministically exercises the micro-compaction path regardless of
the environment.
* fix(agent-core): harden full compaction against in-flight races, unbounded shrink, and media loss
Three compaction-path fixes surfaced by review, each flipping its
documenting it.fails probe to a passing it:
- Append race (CMP-02): after the summarizer returns, the post-summary
history check only compared the compacted prefix. A live step appending
to the tail while a manual/SDK compaction was in flight slipped through —
an appended assistant/tool turn is neither summarized (the summary covers
only the snapshot) nor kept (the rebuild keeps user input), so it
vanished. Now cancel when the appended tail contains a non-user message;
an appended user message is still kept (rebuild picks it up), preserving
the existing 'keeps messages appended while compacting an unchanged
prefix' behavior.
- Unbounded empty/truncated shrink: an empty or truncated summary dropped
the oldest message and reset retryCount, so a model that kept returning
empty could issue ~one request per history entry. Bound the shrink
attempts by MAX_COMPACTION_RETRY_ATTEMPTS, mirroring the overflow-shrink
counter.
- Media dropped on truncation (CMP-07): truncating the oldest kept user
message replaced its whole content with one text block, discarding any
image/audio/video. Keep the non-text parts and spend the remaining budget
(maxTokens minus their cost) on truncated text.
* fix(vis): mirror legacy compaction tail in the model-mode projector
For a pre-rework context.apply_compaction record (no keptUserMessageCount),
agent-core's ContextMemory restore and the transcript reducer keep the old
[summary, ...history.slice(compactedCount)] tail — a verbatim recent tail
including assistant/tool. The vis model-mode projector always applied the
new kept-user selection, so opening an older compacted session in model
mode hid the assistant/tool tail the resumed agent still holds (and
surfaced a pre-compaction user message the agent dropped).
Branch on a missing keptUserMessageCount with compactedCount < history
length and reproduce the legacy shape, matching the agent-core restore.
* fix(agent-core): cancel compaction on any droppable user-role tail
The in-flight append guard cancelled only when the tail grew with a
non-user role. A user-role message that compaction would still drop — a
background-task notification, hook/cron reminder, or shell-command output —
slipped through: appended after the summary snapshot (so absent from the
summary) and dropped by the all-user rebuild (which keeps only real user
input), vanishing silently.
Key the guard on the same predicate applyCompaction uses (!isRealUserInput)
so it cancels whenever the appended tail holds anything compaction would
drop. A real user message is still kept, so a live user turn racing a
manual/SDK compaction continues to complete.
* fix(agent-core): exclude pre-clear prompts from legacy folded length
The transcript reducer's legacy fallback (records predating
keptUserMessageCount, compacted with no verbatim tail) re-derived the
kept-user count from the whole transcript, including messages before the
last context.clear. Live ContextMemory rebuilds _history from post-clear
messages only, so counting pre-clear prompts overstated foldedLength;
MessageService then saw context.history.length <= foldedLength and skipped
appending unflushed live tail messages, dropping recent output from the
messages endpoint for old sessions compacted after a clear.
Derive only from entries at or after clearFloor to match the live context.
* fix(agent-core): drop media when truncating the oldest kept prompt
Revert the media-preserving truncation: keeping non-text parts on the
truncated boundary message overshot the kept-user budget when the media
alone exceeded it, and reordered interleaved text/media parts. Both codex
(no media-aware truncation) and Claude Code (strips media at compaction)
decline to preserve media on a truncated message, since media cannot be
partially truncated and keeping it whole breaks the budget.
truncateUserMessage now keeps only the truncated text. Recent messages
that fit the budget are still kept verbatim with their media; only the
oldest, partially-overflowing boundary message loses its attachments.
* fix(agent-core): make manual compaction and turns mutually exclusive
A manual/SDK compaction could start while a turn was streaming, or a new
turn could launch while a compaction was in flight. Either way the turn
mutates the shared context (streaming content into an existing assistant
message, or appending new messages) during the summarizer await, and that
output is neither summarized nor preserved by the all-user rebuild —
silent loss that object-identity checks can't detect (the streamed message
is mutated in place).
Guard both directions so the agent does one of {turn, compaction} at a
time: begin() refuses a manual compaction while a turn is active, and
launch() refuses a new turn while a compaction is in progress. Auto
compaction is exempt — it runs from within the turn at a step boundary,
which blocks the turn for its duration.
* chore(changeset): consolidate compaction changesets into one
* chore(agent-core): drop external-product references from compaction comments
* test(agent-core): add Anthropic wire-compliance smoke tests for compaction
Drive real compaction output and the compaction summarizer projection
through the real Anthropic provider conversion and assert the wire request
is well-formed: strict user/assistant alternation and every tool_use
answered by an adjacent tool_result. Locks in the cross-layer guarantee
(projector merge + Anthropic consecutive-user merge + adjacency repair +
synthesizeMissing) that compacted sessions stay valid for strict
Anthropic-compatible backends.
* fix(agent-core): defer and replay inputs during manual compaction instead of rejecting
Manual/SDK compaction runs outside a turn, so the earlier guard rejected
prompts/steers that arrived while it held the context. That broke three
things: a REST/web prompt got stuck 'running' (no terminal turn event), a
background-task/cron steer was silently lost (null was read as 'buffered'
but nothing was), and a follow-up prompt could land in the window after
isCompacting cleared but before reminders were reinjected.
Reuse the existing defer-and-replay model instead of rejecting:
- steer() and launch() buffer into steerBuffer while a compaction is in
progress (returning null = buffered), mirroring how an active turn defers
input.
- FullCompaction.compactionWorker keeps isCompacting true through
refreshSystemPrompt + injectAfterCompaction (moving markCompleted and the
completed event after reinjection), then replays the buffer via
TurnFlow.onCompactionFinished — on success, on an A1 prefix/tail cancel,
and on failure/abort.
- onCompactionFinished flushes into an active turn if one exists, else
launches a fresh turn from the deferred input.
No PromptService change: a deferred prompt's eventual turn.started lets it
associate the pending prompt and clear it on turn.ended.
* feat(kosong): detect tool_use/tool_result adjacency errors
Add isToolExchangeAdjacencyError to classify the strict-provider 400 raised
when an assistant tool_use is not correctly paired with its tool_result
(missing, stray, or non-adjacent), excluding context-overflow 400s. Lets the
agent loop recognize the error and resend a wire-compliant request instead of
leaving the session stuck.
* fix(agent-core): close mid-history orphan tool calls and resend wire-compliant after a strict 400
Strict providers (Anthropic) reject a request whose assistant tool_use is not
answered by an adjacent tool_result, and the same malformed history is re-sent
every turn, permanently bricking the session.
- Projector now closes a mid-history tool call whose result is missing entirely
(a later turn proves it is not in-flight) with a synthetic result; the
trailing in-flight call is still left untouched.
- Add a strict projection (synthesize every open call, drop stray results) and,
on a tool_use/tool_result adjacency 400, resend the request once with it.
- Report every projection repair (reorder / synthesize / drop) via log and
telemetry, deduped by signature, so a silently-mangled history leaves a trace.
Trailing-tail synthesis (expected under compaction) is not flagged.
* fix(kosong): merge consecutive user turns for strict providers
Gemini/Vertex require strictly alternating user/model turns and reject
consecutive user turns with HTTP 400. They arise after compaction (kept
prompts + user-role summary + injected reminders) and when a turn is
steered in right after a tool result. Anthropic already merged them
inline; the Google converter did not, so post-compaction requests failed.
Extract the asymmetric merge into a shared mergeConsecutiveUserMessages
helper applied at each strict provider's conversion boundary: refactor
Anthropic to use it (behavior unchanged) and apply it at the Google
converter's exit. A conformance suite drives every strict provider with
the post-compaction shape and a steer-after-tool-result shape, asserting
no consecutive same-role turns reach the wire, so a new strict provider
cannot silently omit the merge.
The provider-agnostic projector stays structure-preserving: lenient
providers (OpenAI/Kimi) keep distinct turns for clearer message
boundaries; only strict providers normalize, where the requirement lives.
* feat(kosong): recognize the broader structural request-rejection family
Add isRecoverableRequestStructureError, covering the strict-provider 400s that
stem from a malformed message array re-sent every turn: tool_use/tool_result
pairing, empty/whitespace-only text blocks, a non-user first message, and
non-alternating roles. Context-overflow 400s are excluded (handled by
compaction). Lets the loop trigger one strict, wire-compliant resend for the
whole family rather than only tool-pairing errors.
* fix(agent-core): sanitize whitespace and strict-resend structural 400s, with diagnostics
- Drop empty AND whitespace-only text blocks in projection (Anthropic rejects
whitespace-only with "text content blocks must contain non-whitespace text",
which otherwise sticks a session); treat whitespace-only tool output as empty.
- Broaden the post-400 strict resend to the whole structural family and add two
strict-only passes to the strict projection: drop leading non-user messages
(first message must be user) and merge consecutive assistant turns.
- Log + telemetry for every wire repair the projector applies (reorder,
synthesize, drop orphan, drop leading, merge assistants, drop whitespace),
deduped by signature; log the strict resend outcome (recovered or still
rejected) so a stuck session always leaves a trace.
* fix(agent-core): normalize empty-equivalent tool result arrays to the empty placeholder
A tool result whose ContentPart[] output has no sendable content (an empty array,
or only empty/whitespace-only text blocks) was returned verbatim, so projection
stripped the blank blocks, left the tool message empty, and threw on every send —
bricking the session locally. String outputs were already normalized; do the same
for arrays. A non-text part or any non-whitespace text still keeps the real
output.
* chore(changeset): simplify the wire-compliance changeset
* feat(agent-core): rework compaction to keep only user prompts and summary
* refactor(agent-core): rewrite compaction summary as first-person handoff
Rework the full-compaction summary to read as the agent's own continuing
notes instead of a third-party report:
- compaction-instruction.md: free-form first-person continuation that
preserves exact commands, paths and outcomes, states the precise next
action, and flags claimed-but-unverified work rather than trusting it.
- compaction-summary-prefix.md: skeptical "your own working notes"
framing; drop the collaborative third-party prefix.
- system.md: add compaction-awareness guidance so the model continues
naturally from a summary and re-checks any reported "done".
- Rename the compaction helpers module to handoff.ts.
Update tests and regenerate snapshots for the new prompt text, and fill
in contextSummary in the restored-compaction replay expectations.
* fix(agent-core): count image/audio/video parts in token estimation
estimateTokensForContentPart returned 0 for image_url/audio_url/video_url,
so auto-compaction triggers, the overflow-shrink budget, the kept-user
budget, and the reported context size all went blind to media — a
media-heavy session could overflow the model window while the estimate
reported a near-empty context. Media parts now carry a fixed estimate
(MEDIA_TOKEN_ESTIMATE), and the content-part switch is exhaustive so a new
ContentPart kind must declare its estimate rather than silently count as
zero.
* feat(agent-core): re-surface active background tasks after compaction
Folding the live context to [recent user prompts, summary] drops the
messages that started background tasks and their status updates, so the
model could forget a task is still running and spawn a duplicate.
injectAfterCompaction now appends a system-reminder listing active
background tasks (with guidance to use TaskOutput/TaskList/TaskStop
instead of re-spawning). It runs only post-compaction and carries an
injection origin, so the next compaction drops and rebuilds it rather
than stacking copies; the all-user-role post-compaction shape is
preserved (no tool-pairing reintroduced).
* test(agent-core): add compaction scenario guards and risk probes
Adds compaction-scenarios.test.ts driving the real Agent/ContextMemory/
FullCompaction machinery:
- A guard test locking in that repeated compaction folds the prior summary
into the new one instead of stacking two summaries.
- Seven `it.fails` probes that executably reproduce known, currently-accepted
edge-case defects so the suite stays green while documenting each one
precisely; any of them will flip red (forcing removal of `.fails`) the day
the behavior is fixed. They cover: assistant/tool appended during an
in-flight summarizer call being dropped; unbounded shrink on empty
summaries; the fixed 20k kept-user budget overflowing a small model window;
a tool result orphaned when compaction starts mid-exchange; legacy
compaction records dropping their verbatim tail on replay; micro-compaction
clearing recent tool results in an overflow-shrunk suffix; and media being
discarded when the oldest kept user message is truncated.
* fix(agent-core): repair tool_use/tool_result adjacency in projected context
A tool call and its result can end up non-adjacent in history — a
background-task notification or flushed steer lands between them, or an
interrupted/nested step delays the result — which strict providers reject
with HTTP 400. The projector now moves each tool_use's result up to
immediately follow it (projection-time only; the stored history is
untouched), and full compaction projects its summarizer input with a
synthetic result for any still-open call so the summary request stays
well-formed. Micro-compaction only surfaced this latent ordering by busting
the prompt cache, so it now defaults off.
Includes projector adjacency regression tests, a context-level integration
test, and a compaction synthesize-missing guard; the prior "keeps an
unresolved tool exchange out of the compaction prompt" test is updated to
the now-well-formed (synthetic-result) behavior.
* fix(agent-core): preserve the verbatim tail when restoring legacy compactions
A pre-rework `context.apply_compaction` record used
`[summary, ...history.slice(compactedCount)]` semantics and kept a verbatim
recent tail, but it has no `keptUserMessageCount`. The reworked applyCompaction
re-folded such records into the all-user shape, dropping the recent
assistant/tool tail — so resuming a session compacted by an older version
silently lost its most recent context.
On restore of such a record (gated on records.restoring, no keptUserMessageCount,
and compactedCount < history length) reproduce the old shape instead. The
forward/live path is unchanged; the projector's tool-adjacency repair keeps the
restored tail well-formed, and compaction only runs at clean step boundaries so
the tail has no open exchange. The legacy-tail probe now passes as a regression
guard via the real restore path.
* fix(agent-core): align legacy compaction foldedLength with live restore
The transcript reducer re-derived foldedLength for pre-rework
context.apply_compaction records (no keptUserMessageCount) using the new
kept-user+summary rule, but ContextMemory's restore now reproduces the legacy
[summary, ...history.slice(compactedCount)] shape for those records. The two
diverged for legacy sessions, so MessageService's foldedLength-vs-live-history
comparison could mis-handle GET /messages (miss or misorder recent output).
The reducer now mirrors the live legacy fold: when compactedCount is below the
pre-compaction length it computes 1 + (length - compactedCount); otherwise it
falls back to the kept-user derivation. The MessageService transcript test's
fixture is corrected to a new-format record, matching its all-user live mock.
* fix(kosong): merge a follow-up user turn into the preceding tool_results
The Anthropic message merge keyed on isToolResultOnly(last) ===
isToolResultOnly(converted), which left a tool_result-only user turn
followed by a plain-text user turn unmerged. After tool-exchange repair
this shape (assistant tool_use -> tool_result -> injected notification)
produces two adjacent user messages, which strict Anthropic-compatible
backends reject with HTTP 400.
Switch to the asymmetric predicate isToolResultOnly(last) ||
!isToolResultOnly(converted): a tool-result-only running message absorbs
whatever user turn follows (parallel tool_results or a trailing text),
yielding a valid [tool_result, ..., text] message; a plain-text running
message still only absorbs plain text. [tool_result, text] is valid for
both native Anthropic (which concatenates anyway) and strict backends.
* test(agent-core): pin micro-compaction flag in the shrunk-suffix probe
The 'does not clear recent tool results when projecting a shrunk suffix'
probe is an it.fails that only documents a real defect while
micro-compaction is active. It inherited the ambient
KIMI_CODE_EXPERIMENTAL master switch, so its pass/fail flipped with the
runner: green locally (master switch on) but a hard failure in CI, where
the flag defaults off and MicroCompaction.compact() is a no-op that
leaves the tool result intact.
Enable KIMI_CODE_EXPERIMENTAL_MICRO_COMPACTION explicitly for this probe
so it deterministically exercises the micro-compaction path regardless of
the environment.
* fix(agent-core): harden full compaction against in-flight races, unbounded shrink, and media loss
Three compaction-path fixes surfaced by review, each flipping its
documenting it.fails probe to a passing it:
- Append race (CMP-02): after the summarizer returns, the post-summary
history check only compared the compacted prefix. A live step appending
to the tail while a manual/SDK compaction was in flight slipped through —
an appended assistant/tool turn is neither summarized (the summary covers
only the snapshot) nor kept (the rebuild keeps user input), so it
vanished. Now cancel when the appended tail contains a non-user message;
an appended user message is still kept (rebuild picks it up), preserving
the existing 'keeps messages appended while compacting an unchanged
prefix' behavior.
- Unbounded empty/truncated shrink: an empty or truncated summary dropped
the oldest message and reset retryCount, so a model that kept returning
empty could issue ~one request per history entry. Bound the shrink
attempts by MAX_COMPACTION_RETRY_ATTEMPTS, mirroring the overflow-shrink
counter.
- Media dropped on truncation (CMP-07): truncating the oldest kept user
message replaced its whole content with one text block, discarding any
image/audio/video. Keep the non-text parts and spend the remaining budget
(maxTokens minus their cost) on truncated text.
* fix(vis): mirror legacy compaction tail in the model-mode projector
For a pre-rework context.apply_compaction record (no keptUserMessageCount),
agent-core's ContextMemory restore and the transcript reducer keep the old
[summary, ...history.slice(compactedCount)] tail — a verbatim recent tail
including assistant/tool. The vis model-mode projector always applied the
new kept-user selection, so opening an older compacted session in model
mode hid the assistant/tool tail the resumed agent still holds (and
surfaced a pre-compaction user message the agent dropped).
Branch on a missing keptUserMessageCount with compactedCount < history
length and reproduce the legacy shape, matching the agent-core restore.
* fix(agent-core): cancel compaction on any droppable user-role tail
The in-flight append guard cancelled only when the tail grew with a
non-user role. A user-role message that compaction would still drop — a
background-task notification, hook/cron reminder, or shell-command output —
slipped through: appended after the summary snapshot (so absent from the
summary) and dropped by the all-user rebuild (which keeps only real user
input), vanishing silently.
Key the guard on the same predicate applyCompaction uses (!isRealUserInput)
so it cancels whenever the appended tail holds anything compaction would
drop. A real user message is still kept, so a live user turn racing a
manual/SDK compaction continues to complete.
* fix(agent-core): exclude pre-clear prompts from legacy folded length
The transcript reducer's legacy fallback (records predating
keptUserMessageCount, compacted with no verbatim tail) re-derived the
kept-user count from the whole transcript, including messages before the
last context.clear. Live ContextMemory rebuilds _history from post-clear
messages only, so counting pre-clear prompts overstated foldedLength;
MessageService then saw context.history.length <= foldedLength and skipped
appending unflushed live tail messages, dropping recent output from the
messages endpoint for old sessions compacted after a clear.
Derive only from entries at or after clearFloor to match the live context.
* fix(agent-core): drop media when truncating the oldest kept prompt
Revert the media-preserving truncation: keeping non-text parts on the
truncated boundary message overshot the kept-user budget when the media
alone exceeded it, and reordered interleaved text/media parts. Both codex
(no media-aware truncation) and Claude Code (strips media at compaction)
decline to preserve media on a truncated message, since media cannot be
partially truncated and keeping it whole breaks the budget.
truncateUserMessage now keeps only the truncated text. Recent messages
that fit the budget are still kept verbatim with their media; only the
oldest, partially-overflowing boundary message loses its attachments.
* fix(agent-core): make manual compaction and turns mutually exclusive
A manual/SDK compaction could start while a turn was streaming, or a new
turn could launch while a compaction was in flight. Either way the turn
mutates the shared context (streaming content into an existing assistant
message, or appending new messages) during the summarizer await, and that
output is neither summarized nor preserved by the all-user rebuild —
silent loss that object-identity checks can't detect (the streamed message
is mutated in place).
Guard both directions so the agent does one of {turn, compaction} at a
time: begin() refuses a manual compaction while a turn is active, and
launch() refuses a new turn while a compaction is in progress. Auto
compaction is exempt — it runs from within the turn at a step boundary,
which blocks the turn for its duration.
* chore(changeset): consolidate compaction changesets into one
* chore(agent-core): drop external-product references from compaction comments
* test(agent-core): add Anthropic wire-compliance smoke tests for compaction
Drive real compaction output and the compaction summarizer projection
through the real Anthropic provider conversion and assert the wire request
is well-formed: strict user/assistant alternation and every tool_use
answered by an adjacent tool_result. Locks in the cross-layer guarantee
(projector merge + Anthropic consecutive-user merge + adjacency repair +
synthesizeMissing) that compacted sessions stay valid for strict
Anthropic-compatible backends.
* fix(agent-core): defer and replay inputs during manual compaction instead of rejecting
Manual/SDK compaction runs outside a turn, so the earlier guard rejected
prompts/steers that arrived while it held the context. That broke three
things: a REST/web prompt got stuck 'running' (no terminal turn event), a
background-task/cron steer was silently lost (null was read as 'buffered'
but nothing was), and a follow-up prompt could land in the window after
isCompacting cleared but before reminders were reinjected.
Reuse the existing defer-and-replay model instead of rejecting:
- steer() and launch() buffer into steerBuffer while a compaction is in
progress (returning null = buffered), mirroring how an active turn defers
input.
- FullCompaction.compactionWorker keeps isCompacting true through
refreshSystemPrompt + injectAfterCompaction (moving markCompleted and the
completed event after reinjection), then replays the buffer via
TurnFlow.onCompactionFinished — on success, on an A1 prefix/tail cancel,
and on failure/abort.
- onCompactionFinished flushes into an active turn if one exists, else
launches a fresh turn from the deferred input.
No PromptService change: a deferred prompt's eventual turn.started lets it
associate the pending prompt and clear it on turn.ended.
* fix(kosong): merge consecutive user turns for strict providers
Gemini/Vertex require strictly alternating user/model turns and reject
consecutive user turns with HTTP 400. They arise after compaction (kept
prompts + user-role summary + injected reminders) and when a turn is
steered in right after a tool result. Anthropic already merged them
inline; the Google converter did not, so post-compaction requests failed.
Extract the asymmetric merge into a shared mergeConsecutiveUserMessages
helper applied at each strict provider's conversion boundary: refactor
Anthropic to use it (behavior unchanged) and apply it at the Google
converter's exit. A conformance suite drives every strict provider with
the post-compaction shape and a steer-after-tool-result shape, asserting
no consecutive same-role turns reach the wire, so a new strict provider
cannot silently omit the merge.
The provider-agnostic projector stays structure-preserving: lenient
providers (OpenAI/Kimi) keep distinct turns for clearer message
boundaries; only strict providers normalize, where the requirement lives.
* feat: support multi-level thinking effort switching
- kimi provider: emit thinking.effort in the new wire format; keep reasoning_effort mirrored during the transition
- model catalog: thread support_efforts / default_effort from oauth through to /models
- config schema: add supportEfforts / defaultEffort on model aliases
- TUI: multi-segment thinking control in /model, new /effort command, footer effort display
- switch status uses displayName and distinguishes model vs effort-only changes
* docs: add thinking effort design plans
- thinking-effort-switching.md: implemented multi-level effort switching
- thinking-model-overhaul.md: follow-up refactor plan for the thinking state model
* docs: collapse thinking overhaul plan into a single PR
* refactor!: overhaul thinking config and effort resolution
Replace default_thinking and thinking.mode with a single [thinking] enabled/effort table. ThinkingEffort is now an open string ('off' | 'on' | model-declared effort); effort levels come from each model's support_efforts instead of a fixed enum.
Centralize default and always_thinking clamp logic in resolveThinkingEffort/defaultThinkingEffortFor, and honor an explicitly configured effort when an always_thinking model is forced back on.
TUI keeps a single thinkingEffort field instead of the boolean + level pair; 'on' is normalized to the model default at the UI boundary.
BREAKING CHANGE: default_thinking and thinking.mode are removed from config; migrate to [thinking] enabled/effort.
* refactor: rename residual thinking level wording to effort
Rename comments, error messages, parameter names, the SetThinkingPayload wire field (level -> effort), and TUI local variables so the thinking effort naming is consistent throughout. No behavior change.
* refactor: rename remaining camelCase thinking level identifiers to effort
Rename liveLevel/prevLevel/levelChanged/commitLevel/effectiveLevel to liveEffort/prevEffort/effortChanged/commitEffort/effectiveEffort in the TUI model picker and config commands.
* refactor: eliminate remaining thinking level wording in comments and tests
Rename levelLabel -> effortLabel, EffortSelectorOptions.levels -> efforts, and 'effort level(s)' / 'default level' / 'requested level' wording in comments, error messages, slash-command description, and test titles to effort. Also restore the withThinking(effort) parameter rename in the Kimi provider that was accidentally reverted.
* fix: address codex review feedback on thinking effort handling
- OpenAI thinkingEffortToReasoningEffort and Anthropic clampEffort now normalize 'on' / unrecognized efforts instead of throwing, so boolean non-Kimi models no longer crash on session start.
- ACP resolveCurrentThinkingEnabled treats a non-empty thinking.effort as enabled, matching agent-core's resolveThinkingEffort.
- REST promptThinkingSchema accepts any non-empty effort string so model-declared efforts are not rejected at the API boundary.
* test: align kimi e2e expectations with supportEfforts-gated reasoning_effort
The kimi provider now sends reasoning_effort only when the model declares support_efforts; boolean models (no support_efforts) send only thinking.type. Update the kimi e2e tests to drop the stale reasoning_effort expectation for the boolean test model.
* test: cover [thinking] effort parsing in config.test
Add effort = "high" to the documented [thinking] table in the config parse test and assert config.thinking.effort is resolved, so the new [thinking] effort field has direct parse coverage.
* docs: add thinking test coverage gap analysis
Capture the explore agent's test coverage review for the thinking overhaul PR, including P1/P2 gaps and the two open design questions, for follow-up test additions.
* feat(oauth): parse nested think_efforts from /models response
The /models endpoint now returns effort levels under a nested think_efforts object ({ support, valid_efforts, default_effort }). Parse it preferentially in both managed-kimi-code and open-platform model parsing, falling back to the legacy flat support_efforts / default_effort fields for older servers.
* refactor(oauth): only read nested think_efforts; gate on support=true
Drop the legacy flat support_efforts / default_effort fallback. The think_efforts object is now the single source, and its support flag gates the whole object — when support is not true, valid_efforts and default_effort are ignored entirely.
* chore: remove unused parseStringArray import in open-platform
* docs: finalize thinking effort release notes
Downgrade the changeset to minor with an English summary, drop the version-specific 'added in 1.0.0' info block, and present the deprecated config fields as a table (field / deprecated in 0.21.0 / description).
* refactor: drop temporary refresh toggles and kimi reasoning_effort mirror
Remove the always-true REFRESH_MODELS_ON_PICKER_OPEN / REFRESH_PROVIDER_MODELS_ON_STARTUP toggles and their stale re-enable TODOs, and stop sending reasoning_effort from the kimi provider (thinking.effort is the only wire field now).
* fix(tui): avoid persisting "on" as thinking effort
* fix: preserve persisted thinking effort across login and provider setup
* fix(tui): show actual thinking effort in /status and footer
* test(tui): align message-flow expectations with effort persistence and /status display
* fix(vis): rename thinkingLevel to thinkingEffort in config.update analysis
* feat(timing): split TTFT into api-server and client portions
Time-to-first-token previously lumped in-process request building
(message serialization, param assembly) together with network + server
latency, making it impossible to tell whether a slow turn was the client
or the API server.
Add an `onRequestSent` hook to kosong's GenerateOptions, fired by every
provider immediately before it dispatches the network call. The window
from request start to dispatch is attributed to the client; the window
from dispatch to the first streamed token is attributed to the API
server. The split flows through the step.end / turn.step.completed
events (and therefore wire.jsonl) and is surfaced in three places:
- KIMI_CODE_DEBUG=1: `TTFT: 2.5s (api 2.4s + client 100ms)`
- session log: new `llm response` line with the timing breakdown
- vis: firstToken/api + firstToken/client rows and timeline label
The split is omitted (total only) when a provider does not report the
boundary, preserving backward compatibility.
* feat(timing): split the decode window into server vs client time
Time-to-first-token now reports a client/server split, but the slow part
of a long turn is the decode window (inter-token streaming), which was
still a single opaque number. Profiling long sessions showed decode
throughput halving over a session's lifetime independent of context
size, which the synchronous per-chunk stream pipeline can cause: kosong
awaits the host callback for every streamed part, so a loaded main
thread throttles how fast tokens are pulled off the wire.
Account for this directly in the stream loop: the time awaiting the next
part (server + network) versus the time spent processing each part
in-process (deep copy, host callback, part merge). The split is reported
through onStreamEnd and flows through the step.end / turn.step.completed
events (and wire.jsonl) into the same three surfaces as the TTFT split:
- KIMI_CODE_DEBUG=1: `TPS: 40.0 tok/s (200 tokens in 5.0s; server 4.6s + client 400ms)`
- session log: serverDecodeMs / clientConsumeMs on the `llm response` line
- vis: streamDuration/server + streamDuration/client rows and timeline label
A large, growing client share confirms host-side throttling; a dominant
server share points at the server/connection. The per-chunk accounting is
wrapped in try/finally so it stays correct across `continue` and aborts,
and is omitted when the stream reports nothing.
* feat(managed-kimi-code): route anthropic protocol via beta api
- kosong: add betaApi option to use client.beta.messages.create
- agent-core: thread alias betaApi into the anthropic provider config
- oauth: route managed models on the anthropic protocol through the beta Messages API
* feat(providers): add KIMI_CODE_CUSTOM_HEADERS support
- Add KIMI_CODE_CUSTOM_HEADERS env var for custom outbound LLM headers
- Send User-Agent to non-Kimi providers
- Forward Kimi identity headers to model catalog fetches
- Support defaultHeaders in Google GenAI provider
* feat(agent-core): add protocol attrs to turn and api error telemetry
- Add type/protocol/alias to api_error for per-protocol error attribution
- Add turn_ended event with reason/duration/mode/type/protocol
- Add type/protocol to turn_interrupted
* chore(oauth): remove hardcoded internal dev endpoint from shared OAuth base URLs
---------
Co-authored-by: haozhe.yang <yanghaozhe@moonshot.ai>
* fix(agent-core): recover from context overflow 413
- track provider-observed effective context limit after overflow
- compact with the reduced limit before retrying the turn
- treat large plain 413 responses as recoverable context overflow
- add CLI patch changeset
* feat(managed-kimi-code): support Anthropic-compatible protocol
- switch managed provider to anthropic when models declare anthropic protocol
- add base64 video content blocks to the kosong anthropic provider
- downgrade unsupported media parts to text placeholders by capability
- pass prompt cache key as Anthropic metadata.user_id for session affinity
* feat(agent-core): add protocol/type to request and video upload telemetry
- turn_started now carries `type` (configured provider wire type) and
`protocol` (effective transport, i.e. alias.protocol ?? provider.type)
- new video_upload event reports mime type, size, latency and
success/failure, plus type/protocol/model context
- ResolvedRuntimeProvider gains `type` and `protocol` fields
* feat: cap completion tokens to remaining context window for chat-completions
* test: cover dynamic completion budget for kimi and openai-legacy
* fix: leave compaction budget uncapped to avoid one-token summaries
* fix: add hookCount to plugins-selector test mocks
* 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.
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(kosong): support claude-fable-5 adaptive thinking
claude-fable-5 only accepts thinking: {type: "adaptive"} with
output_config.effort; the legacy enabled/budget_tokens config and an
explicit disabled config both return HTTP 400.
- Parse the fable family in Claude model ids (major-only version)
- Route fable >= 5 to adaptive thinking; allow xhigh effort
- Omit the thinking field entirely when thinking is off on fable
- Register the 128k output ceiling and thinking/vision capability
* update .changeset
Updated the configuration to use adaptive thinking for Claude Fable 5 support.
Signed-off-by: 7Sageer <12210216@mail.sustech.edu.cn>
---------
Signed-off-by: 7Sageer <12210216@mail.sustech.edu.cn>
* feat(env): migrate kimi-cli model request params and auto-update toggle
Migrate still-relevant environment variables from kimi-cli:
- KIMI_MODEL_TEMPERATURE / KIMI_MODEL_TOP_P: sampling params applied
globally to any kimi provider (not tied to KIMI_MODEL_NAME).
- KIMI_MODEL_THINKING_KEEP: Moonshot preserved-thinking passthrough
(thinking.keep), injected only while Thinking is on.
- KIMI_CODE_NO_AUTO_UPDATE (legacy alias KIMI_CLI_NO_AUTO_UPDATE):
fully disables the update preflight.
Wires env -> provider in Agent.get llm() via applyKimiEnvGenerationParams,
reusing kosong's existing GenerationKwargs / thinking.keep support.
KIMI_MODEL_MAX_TOKENS is intentionally untouched: it already flows through
the completion-budget path.
* fix(env): apply Kimi sampling params to compaction requests too
Sink KIMI_MODEL_TEMPERATURE / KIMI_MODEL_TOP_P into ConfigState.provider so
every request built from config.provider — main loop and full-history
compaction alike — carries them, matching kimi-cli where these live on the
shared create_llm provider. thinking.keep stays in Agent.llm because it
depends on the runtime thinking state (compaction runs thinking-off and
correctly skips it).
Splits applyKimiEnvGenerationParams into applyKimiEnvSamplingParams (applied
at provider construction) and applyKimiEnvThinkingKeep (applied in Agent.llm).
Addresses PR review feedback about compaction requests bypassing the wrapped
provider.
