@oh-my-pi/pi-ai 18.12.1 → 18.12.3

package/CHANGELOG.md

@@ -2,6 +2,33 @@
 
 ## [Unreleased]
 
+## [18.12.3] - 2026-10-19
+
+### Added
+
+- Added `validateToolArguments(tool, call, { keepUnknownRootKeys: true })` to keep unknown root keys that a closed schema would otherwise drop during argument repair ([#16317](https://github.com/can1357/oh-my-pi/issues/16317))
+
+### Fixed
+
+- Fixed Claude streams that a gateway cut off before `message_delta` failing the turn instead of being retried ([#16341](https://github.com/can1357/oh-my-pi/pull/16341) by [@lz-lunzi](https://github.com/lz-lunzi))
+- Fixed Gemini pairing parallel same-name tool results with the wrong call when history stores them out of call order (sessions from before 18.4.2, SDK-built contexts) ([#16259](https://github.com/can1357/oh-my-pi/pull/16259) by [@will-bogusz](https://github.com/will-bogusz))
+- Fixed Factory Droid Gemini dropping calls to tools with names over 64 characters (most MCP tools) and their results from later requests, so the model called the tool again; such names now shorten to 64 characters ([#16283](https://github.com/can1357/oh-my-pi/pull/16283) by [@will-bogusz](https://github.com/will-bogusz))
+- Fixed Factory Droid Gemini requests failing with 400 "For schema with items, schema type should be ARRAY" when a tool parameter accepts a string or a list, as the default `task` tool's `model` field does ([#16249](https://github.com/can1357/oh-my-pi/pull/16249) by [@will-bogusz](https://github.com/will-bogusz))
+- Fixed `PI_REQ_DEBUG=1` request dumps writing your API key, bearer token and response cookies to disk in plain text: credential headers now read `[redacted]`, as in rejected-request dumps ([#16293](https://github.com/can1357/oh-my-pi/pull/16293) by [@will-bogusz](https://github.com/will-bogusz))
+- Fixed finalized signed reasoning disappearing from displayed thinking when no deltas streamed ([#16227](https://github.com/can1357/oh-my-pi/issues/16227))
+
+## [18.12.2] - 2026-10-18
+
+### Added
+
+- Added `nativeJudgmentImageBounds()`: the image bounds a native judge enforces, its wire's transport limits tightened by the catalog's model bounds
+- Added Box judgment questions (`BoxQuestion`/`BoxAnswer`, TypeSafe Jev 1.14+ direct or through OpenRouter): each names the request image it searches by `image` index, and comes back as that image's edges as 0–1 fractions; text, OpenAI Decisions, and Workers AI judges reject them
+
+### Fixed
+
+- Fixed Muse Code usage going stale whenever its 5-hour window sits idle: the idle window now reports as unused, and weekly usage carries over from the last active window until its reset ([#16193](https://github.com/can1357/oh-my-pi/pull/16193) by [@eggpeat](https://github.com/eggpeat))
+- Fixed DeepSeek tool calls never running when the model writes them as chat-template tokens with `_` in place of `▁`, a missing `｜`, or no per-call close (e.g. `<｜tool_calls_begin｜>…<｜tool_calls_end｜>`): they now run as tool calls, and any that still cannot be parsed are removed from history and reported to the model as a failed call ([#16216](https://github.com/can1357/oh-my-pi/pull/16216) by [@H4vC](https://github.com/H4vC))
+
 ## [18.12.1] - 2026-10-17
 
 ### Fixed
@@ -2320,26 +2347,4 @@
 - Fixed OpenAI Responses native history replay leaking image generation provider-only fields into the next request, which made OpenAI-compatible proxies reject `pi` tool-calling sessions with `Unknown parameter: input[1].action`. ([#3201](https://github.com/can1357/oh-my-pi/issues/3201))
 - Fixed a stream thought-leakage issue for `gemini-3.5-flash` where the model's internal reasoning JSON could leak into the visible text stream. The stream parser now uses a brace-balanced counting algorithm to accurately slice and discard the leading thought JSON block, with a robust fallback for unescaped double quotes, dynamic tool-name derivation, and preservation of subsequent text deltas without triggering empty-response retries.
 
-## [16.1.10] - 2026-06-21
-
-### Changed
-
-- Improved JSON robustness by replacing external dependency with a custom, high-performance parser
-- Strengthened streaming JSON parsing to prevent non-finite numbers from surfacing as `undefined/NaN`
-- Configured JSON parser to reject JS-specific `NaN` and `Infinity` values for tool arguments
-- Replaced the JSON repair/parse helpers (`parseJsonWithRepair`, `parseStreamingJson`) with a single from-scratch tolerant parser (`RelaxedJson`) that accepts single-quoted strings, unquoted object keys, trailing/stray commas, `//` and `/* */` comments, Python `True`/`False`/`None`, raw control characters, invalid escapes, and unescaped apostrophes (`'it's'`). Final parsing still throws on truncated/garbage input (so a malformed tool call is skipped rather than executed with half-formed args) and rejects JS-only `NaN`/`Infinity`; streaming parsing stays non-throwing and rolls back incomplete trailing tokens instead of surfacing `undefined`/`NaN`. The Cursor provider's ad-hoc regex + JSON5 tool-argument parser now routes through the shared parser.
-
-### Fixed
-
-- Fixed tool call ID normalization for Anthropic-compatible models
-- Fixed Anthropic Messages replay sanitizing malformed tool-call IDs, including aborted native tool calls with empty IDs, so retries no longer send invalid `tool_use.id` / `tool_result.tool_use_id` pairs.
-- Fixed the Codex Responses WebSocket transport attributing a prior turn's output to the current one on a reused connection: a trailing/duplicate frame from a cleanly-completed previous response that slipped past the queue drain could be consumed as this request's terminal (ending the turn with empty output) or as a stale tool call. Frames are now keyed by `response.id` — a frame carrying the previous response's id is dropped, and one carrying a third id (or a regressed `sequence_number`) fails closed so the turn retries instead of mixing two responses' streams. Idless frames (deltas, the rate-limit/metadata preamble, `response.created`-less streams) still pass through, matching upstream codex-rs.
-- Fixed `transformMessages` pulling an earlier, orphaned tool result onto a later tool call that reused the same id (left behind when compaction folded the originating `tool_use` into a summary). The pending-call flush now pairs each call with a result positioned _after_ its assistant turn, so a reused id surfaces its own output rather than a prior turn's.
-- Fixed DashScope 429 rate-limit messages that mention authorization being classified as credential failures, preventing valid API keys from being invalidated after throttling. ([#3172](https://github.com/can1357/oh-my-pi/issues/3172))
-- Fixed OpenCode Go `401 Insufficient balance` quota errors being treated as unknown failures instead of usage-limit errors, restoring credential rotation and fallback chains. ([#3169](https://github.com/can1357/oh-my-pi/issues/3169))
-
-### Removed
-
-- Removed the `partial-json` dependency; streaming JSON parsing now uses the in-house `RelaxedJson` parser.
-
-Older entries are archived in [packages/ai/CHANGELOG.md@bb602d67c9b6](https://github.com/can1357/oh-my-pi/blob/bb602d67c9b685535bbf8ef1a1cc9de3c944ede4/packages/ai/CHANGELOG.md).
+Older entries are archived in [packages/ai/CHANGELOG.md@5eb78e90fa4e](https://github.com/can1357/oh-my-pi/blob/5eb78e90fa4ec4318b9520302f1977fa58c48745/packages/ai/CHANGELOG.md).

package/dist/types/dialect/deepseek.d.ts

@@ -4,6 +4,19 @@ export declare const DEEPSEEK_TOOL_CALLS_END = "<\uFF5Ctool\u2581calls\u2581end\
 export declare const DEEPSEEK_TOOL_CALL_BEGIN = "<\uFF5Ctool\u2581call\u2581begin\uFF5C>";
 export declare const DEEPSEEK_TOOL_CALL_END = "<\uFF5Ctool\u2581call\u2581end\uFF5C>";
 export declare const DEEPSEEK_TOOL_SEPARATOR = "<\uFF5Ctool\u2581sep\uFF5C>";
+/** DeepSeek template tokens that structure a tool call. */
+export declare const DEEPSEEK_TOOL_CALL_TOKENS: Readonly<Record<string, true>>;
+/** A token found in text: its canonical spelling and the text actually written. */
+export interface DeepSeekTokenMatch {
+    readonly token: string;
+    readonly raw: string;
+    readonly index: number;
+}
+/**
+ * Find the earliest of `tokens` in `text` at or after `from`, in its canonical
+ * spelling or any re-spelling described at {@link TEMPLATE_TOKEN_SPELLING}.
+ */
+export declare function findDeepSeekTemplateToken(text: string, tokens: Readonly<Record<string, true>>, from?: number): DeepSeekTokenMatch | undefined;
 export declare class DeepSeekInbandScanner implements InbandScanner {
     #private;
     constructor(options?: InbandScannerOptions);

package/dist/types/judgment/native.d.ts

@@ -41,10 +41,15 @@ export type JudgmentApi = keyof typeof JUDGMENT_WIRES;
 export declare function isJudgmentApi(api: Api): api is JudgmentApi;
 /** Whether both the native protocol and catalog model support image input. */
 export declare function nativeJudgmentSupportsImages(model: Model): boolean;
+/** Whether `model`'s native judgment takes images and its wire forwards box questions. The model itself may still refuse them (Box needs Jev 1.14+). */
+export declare function nativeJudgmentSupportsBoxes(model: Model): boolean;
 /**
- * Whether `model`'s native judgment API takes these exact images: image support plus every documented
- * transport and catalog bound. A judge chain skips a candidate that would refuse them instead of failing on it.
+ * Every image bound `model`'s native judgment enforces: its wire's transport limits tightened by the
+ * catalog's model bounds. `undefined` when it takes no images. Native APIs reject images outside these
+ * rather than resizing them, so callers size images to them first.
  */
+export declare function nativeJudgmentImageBounds(model: Model): JudgmentImageInput | undefined;
+/** Whether `model`'s native judgment API takes these exact images: image support within every {@link nativeJudgmentImageBounds}. */
 export declare function nativeJudgmentAcceptsImages(model: Model, images: readonly ImageContent[]): boolean;
 /**
  * Reject images that `model`'s native judgment API cannot take — no image support, an unsupported or
@@ -52,6 +57,12 @@ export declare function nativeJudgmentAcceptsImages(model: Model, images: readon
  * @throws {AIError.ValidationError} naming the reason.
  */
 export declare function checkNativeJudgmentImages(model: Model, request: JudgmentRequest, imageAssets?: JudgmentImageAssets): void;
+/**
+ * Reject box questions whose `image` is not an index into `request.images`: nothing would say which
+ * image their edges measure.
+ * @throws {AIError.ValidationError} naming the question.
+ */
+export declare function checkBoxQuestions(request: JudgmentRequest): void;
 /** `TYPESAFE_BASE_URL` when set, else the public API root; trailing slashes stripped. */
 export declare function typesafeBaseUrl(): string;
 /** `TYPESAFE_DEFAULT_MODEL` when set, else {@link TYPESAFE_DEFAULT_MODEL}. */

package/dist/types/judgment/types.d.ts

@@ -3,8 +3,9 @@
  *
  * A {@link Judge} answers a map of named questions — {@link ChoiceQuestion}
  * (one option from a fixed set), {@link NoulQuestion} (probability that a
- * yes/no condition holds), {@link ScoreQuestion} (position on ordered levels)
- * — about one {@link JudgmentState}. Every question in a request sees the same
+ * yes/no condition holds), {@link ScoreQuestion} (position on ordered levels),
+ * {@link BoxQuestion} (where a target is in one of the request's images, native
+ * judges only) — about one {@link JudgmentState}. Every question in a request sees the same
  * state and is answered independently, so callers batch independent questions
  * into one call. The shape mirrors TypeSafe's System One API so the native
  * backend ({@link NativeJudge}) forwards requests verbatim, while the text
@@ -41,7 +42,19 @@ export interface ScoreQuestion {
     instructions: string;
     criteria: readonly [string, string, ...string[]];
 }
-export type Question = ChoiceQuestion | NoulQuestion | ScoreQuestion;
+/**
+ * Where one target is in an image (TypeSafe Box, Jev 1.14+): https://docs.typesafe.ai/primitives/box.
+ * Only native judges answer it. `instructions` must name the target and, among several images, which
+ * one to search; `image` stays local and says which entry of {@link JudgmentRequest.images} the
+ * answer's edges measure, so a judge that resizes or pads that image can map them back.
+ */
+export interface BoxQuestion {
+    type: "box";
+    instructions: string;
+    /** Index into the request's `images` of the image searched. */
+    image: number;
+}
+export type Question = ChoiceQuestion | NoulQuestion | ScoreQuestion | BoxQuestion;
 /** Questions keyed by caller-chosen ids; answers come back under the same ids. */
 export type Questions = Record<string, Question>;
 export interface ChoiceAnswer<L extends string = string> {
@@ -67,9 +80,23 @@ export interface ScoreAnswer {
     /** Concentration of the distribution, 0–1. */
     confidence: number;
 }
-export type Answer = ChoiceAnswer | NoulAnswer | ScoreAnswer;
+/** A rectangle's edges as 0–1 fractions of an image's width (`left` < `right`) and height (`top` < `bottom`), from its top-left corner. */
+export interface BoxEdges {
+    top: number;
+    bottom: number;
+    left: number;
+    right: number;
+}
+export interface BoxAnswer {
+    type: "box";
+    /** Smallest rectangle containing the target, on the image the question's `image` names. Always present: absence is a separate yes/no question. */
+    box: BoxEdges;
+    /** How sure the model is of the box, 0–1. */
+    confidence: number;
+}
+export type Answer = ChoiceAnswer | NoulAnswer | ScoreAnswer | BoxAnswer;
 /** The answer type for a question, preserving choice labels. */
-export type AnswerFor<Q extends Question> = Q extends ChoiceQuestion<infer L> ? ChoiceAnswer<L> : Q extends NoulQuestion ? NoulAnswer : ScoreAnswer;
+export type AnswerFor<Q extends Question> = Q extends ChoiceQuestion<infer L> ? ChoiceAnswer<L> : Q extends NoulQuestion ? NoulAnswer : Q extends BoxQuestion ? BoxAnswer : ScoreAnswer;
 export interface JudgmentRequest<Q extends Questions = Questions> {
     state: JudgmentState;
     questions: Q;

package/dist/types/judgment/wire.d.ts

@@ -42,6 +42,8 @@ export interface JudgmentWire {
     imagesRequireOptIn?: boolean;
     /** Whether this wire drops per-image detail metadata. */
     ignoresImageDetail?: boolean;
+    /** Whether this dialect forwards `box` questions; others reject them before sending. */
+    boxes?: boolean;
     /** The response's answers and usage; `undefined` when the payload is not this API's response shape. */
     decode(payload: unknown): DecodedJudgment | undefined;
 }

package/dist/types/providers/google-shared.d.ts

@@ -76,6 +76,23 @@ export declare function isThinkingPart(part: Pick<Part, "thought" | "thoughtSign
  */
 export declare function retainThoughtSignature(existing: string | undefined, incoming: string | undefined): string | undefined;
 export declare const SKIP_THOUGHT_SIGNATURE = "skip_thought_signature_validator";
+/**
+ * Gemini pairs id-less function responses with calls by position. The agent loop stores
+ * parallel results in call order, but histories can still arrive out of call order (sessions
+ * saved by older versions, contexts built through the SDK), so the converters guard the wire:
+ * record each call as it is emitted and each response with its call, then {@link apply} sorts
+ * every turn's matched responses back into call order within the slots they already occupy.
+ */
+export declare class FunctionCallOrder {
+    #private;
+    call(id: string): void;
+    /** The call ordinal of `id`, if that call was recorded. */
+    ordinal(id: string): number | undefined;
+    response(part: Part, callId: string): void;
+    apply(contents: readonly {
+        parts?: Part[];
+    }[]): void;
+}
 /**
  * Convert internal messages to Gemini Content[] format.
  */

package/dist/types/providers/ollama.d.ts

@@ -1,8 +1,13 @@
-import type { StreamFunction, StreamOptions, ToolChoice } from "../types.js";
+import type { Model, StreamFunction, StreamOptions, ToolChoice } from "../types.js";
 export interface OllamaChatOptions extends StreamOptions {
     reasoning?: "minimal" | "low" | "medium" | "high" | "xhigh" | "max";
     disableReasoning?: boolean;
     toolChoice?: ToolChoice;
 }
+/** Whether Ollama chat removes assistant thinking from replayed history. */
+export declare function ollamaChatDropsOwnThinking(model: {
+    api?: string;
+    compat?: Model["compat"];
+} | null | undefined): boolean;
 /** Retry EOS-only Ollama completions before the agent loop sees an empty stop. */
 export declare const streamOllama: StreamFunction<"ollama-chat">;

package/dist/types/providers/transform-messages.d.ts

@@ -52,6 +52,8 @@ export declare function shouldDropTruncatedThinkingOnlyAssistant(msg: AssistantM
  */
 export declare function isSurvivingAssistant(msg: AssistantMessage, model?: Model): boolean;
 export declare function isAnthropicMessagesModel(model: Model): model is Model<"anthropic-messages">;
+/** Whether `turn` came from `model` itself: its own turns replay their native reasoning unchanged. */
+export declare function isSameModelTurn(turn: Pick<AssistantMessage, "provider" | "api" | "model">, model: Pick<Model, "provider" | "api" | "id">): boolean;
 /**
  * Normalize tool call ID for cross-provider compatibility.
  * OpenAI Responses API generates IDs that are 450+ chars with special characters like `|`.

package/dist/types/usage/muse-code.d.ts

@@ -1,2 +1,2 @@
-import type { UsageProvider } from "../usage.js";
+import { type UsageProvider } from "../usage.js";
 export declare const museCodeUsageProvider: UsageProvider;

package/dist/types/utils/dsml-leak.d.ts

@@ -1,25 +1,27 @@
 /**
- * Removal of DeepSeek DSML tool-call markup left in visible assistant text.
+ * Removal of DeepSeek tool-call markup left in visible assistant text.
  *
  * DeepSeek models on OpenAI-compatible hosts sometimes emit their tool calls as
- * raw `<｜DSML｜…>` markup in `content`. A well-formed envelope is healed into a
- * structured call by `StreamMarkupHealing`, but a degraded model can drop the
- * opening `<｜DSML｜tool_calls>` / `<｜DSML｜invoke name="…">` tags (or the host
- * runs no healer at all). That markup must not stay in history: replaying it
- * teaches the model to keep writing tool calls as text (issue #10556). Callers
- * strip it before the message is committed and tell the model its call failed.
+ * text in `content`: the V4 `<｜DSML｜…>` envelope or the V3 chat-template
+ * tokens (`<｜tool▁calls▁begin｜>` …). A well-formed call is healed into a
+ * structured call by `StreamMarkupHealing`, but a degraded model can drop
+ * required tags (or the host runs no healer at all). That markup must not stay
+ * in history: replaying it teaches the model to keep writing tool calls as text
+ * (issue #10556). Callers strip it before the message is committed and tell the
+ * model its call failed.
  */
 import type { AssistantMessage, Model } from "../types.js";
 /**
  * Whether DSML leak recovery applies to `model`: a DeepSeek-class model, or one
  * explicitly configured with the `dsml` stream healer (e.g. a `models.yml` entry
- * whose id the classifier cannot place). Other models never write DSML tool
- * calls, so DSML text from them is prose and is left alone.
+ * whose id the classifier cannot place). Other models never write DeepSeek tool
+ * calls, so such markup from them is prose and is left alone.
  */
 export declare function isDsmlLeakRecoveryTarget(model: Model): boolean;
 /**
- * Strip leaked DSML tool-call markup from `text`, leaving quoted code intact.
- * Returns `undefined` when `text` carries no such markup.
+ * Strip leaked DeepSeek tool-call markup (DSML or chat-template tokens) from
+ * `text`, leaving quoted code intact. Returns `undefined` when `text` carries
+ * no such markup.
  */
 export declare function stripDsmlToolMarkup(text: string): string | undefined;
 /**

package/dist/types/utils/http-inspector.d.ts

@@ -85,3 +85,5 @@ export declare function rewriteCopilotError(errorMessage: string, error: unknown
  *        credential rotation (sibling keys fail identically).
  */
 export declare function rewriteClinePassError(errorMessage: string, provider: string): string;
+/** Copy of `headers` with every credential value replaced by `[redacted]`; shared by every persisted dump. */
+export declare function redactHeaders<T>(headers: Record<string, T> | undefined): Record<string, T | string> | undefined;

package/dist/types/utils/validation.d.ts

@@ -19,10 +19,20 @@ import type { Tool, ToolCall } from "../types.js";
  * @throws Error if tool is not found or validation fails
  */
 export declare function validateToolCall(tools: Tool[], toolCall: ToolCall): ToolCall["arguments"];
+/** Options for {@link validateToolArguments}. */
+export interface ValidateToolArgumentsOptions {
+    /**
+     * Keep unknown root keys on the validated value instead of letting the
+     * unrecognized-key repair delete them from a closed schema. Set for tools
+     * that own their refusal (`lenientArgValidation`), so execute() can reject
+     * fields the model sent rather than run without them.
+     */
+    keepUnknownRootKeys?: boolean;
+}
 /**
  * Validates tool call arguments against an ArkType or plain JSON Schema schema.
  * Applies conservative LLM-quirk normalization before declaring failure.
  *
  * @throws Error with a formatted message when validation cannot be reconciled.
  */
-export declare function validateToolArguments(tool: Tool, toolCall: ToolCall): ToolCall["arguments"];
+export declare function validateToolArguments(tool: Tool, toolCall: ToolCall, options?: ValidateToolArgumentsOptions): ToolCall["arguments"];

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@oh-my-pi/pi-ai",
-	"version": "18.12.1",
+	"version": "18.12.3",
 	"description": "Unified LLM API with automatic model discovery and provider configuration",
 	"keywords": [
 		"ai",
@@ -155,11 +155,11 @@
 		"fmt": "oxfmt --no-error-on-unmatched-pattern 'src/**/*.{ts,tsx}' '{test,bench,examples,scripts}/**/*.ts' '*.ts'"
 	},
 	"dependencies": {
-		"@oh-my-pi/omptype": "18.12.1",
-		"@oh-my-pi/pi-catalog": "18.12.1",
-		"@oh-my-pi/pi-natives": "18.12.1",
-		"@oh-my-pi/pi-utils": "18.12.1",
-		"@oh-my-pi/pi-wire": "18.12.1"
+		"@oh-my-pi/omptype": "18.12.3",
+		"@oh-my-pi/pi-catalog": "18.12.3",
+		"@oh-my-pi/pi-natives": "18.12.3",
+		"@oh-my-pi/pi-utils": "18.12.3",
+		"@oh-my-pi/pi-wire": "18.12.3"
 	},
 	"devDependencies": {
 		"@types/bun": "^1.3.14"