@tangle-network/agent-runtime 0.85.0 → 0.87.0

package/dist/{chunk-QSYPMMWS.js → chunk-FLVVVBWV.js}

@@ -5,13 +5,13 @@ import {
   definePersona,
   runPersonified,
   worktreeFanout
-} from "./chunk-DLAEEF26.js";
+} from "./chunk-VKVNDNG4.js";
 import {
   createExecutorRegistry
-} from "./chunk-XN5TIJGP.js";
+} from "./chunk-6XCW3M7W.js";
 import {
   runAnalystLoop
-} from "./chunk-YPA5MLVE.js";
+} from "./chunk-LCXXIL3U.js";
 import {
   ConfigError
 } from "./chunk-YEJR7IXO.js";
@@ -196,4 +196,4 @@ export {
   runLoopRunnerCli,
   parseLoopRunnerArgv
 };
-//# sourceMappingURL=chunk-QSYPMMWS.js.map
+//# sourceMappingURL=chunk-FLVVVBWV.js.map

package/dist/{chunk-YPA5MLVE.js → chunk-LCXXIL3U.js}

@@ -1,6 +1,6 @@
 import {
   extractLlmCallEvent
-} from "./chunk-4J6RBI3K.js";
+} from "./chunk-QUXZBZCS.js";
 import {
   AnalystError
 } from "./chunk-YEJR7IXO.js";
@@ -579,4 +579,4 @@ export {
   iterationsToTraceStore,
   runAnalystLoop
 };
-//# sourceMappingURL=chunk-YPA5MLVE.js.map
+//# sourceMappingURL=chunk-LCXXIL3U.js.map

package/dist/{chunk-G7HQI6DB.js → chunk-MHYF2LIP.js}

@@ -8,7 +8,7 @@ import {
   DELEGATION_STATUS_DESCRIPTION,
   DELEGATION_STATUS_INPUT_SCHEMA,
   DELEGATION_STATUS_TOOL_NAME
-} from "./chunk-XN5TIJGP.js";
+} from "./chunk-6XCW3M7W.js";
 
 // src/mcp/openai-tools.ts
 function buildTool(name, description, parameters) {
@@ -45,4 +45,4 @@ export {
   mcpToolsForRuntimeMcp,
   mcpToolsForRuntimeMcpSubset
 };
-//# sourceMappingURL=chunk-G7HQI6DB.js.map
+//# sourceMappingURL=chunk-MHYF2LIP.js.map

package/dist/{chunk-4J6RBI3K.js → chunk-QUXZBZCS.js}

@@ -66,6 +66,83 @@ function pickFiniteNumber(data, keys) {
   }
   return void 0;
 }
+function createSandboxToolPartState() {
+  return { statusByCall: /* @__PURE__ */ new Map(), seq: 0 };
+}
+var TERMINAL_TOOL_FAILURE = /^(error|errored|failed|failure|cancelled|canceled|timeout|timed_out)$/i;
+function mapSandboxToolEvent(event, state) {
+  if (!event || typeof event !== "object") return [];
+  const type = String(event.type ?? "");
+  const data = event.data && typeof event.data === "object" ? event.data : {};
+  if (type === "message.part.updated") {
+    const part = data.part && typeof data.part === "object" ? data.part : {};
+    if (String(part.type ?? "") !== "tool") return [];
+    return projectToolPart(part, state, typeof event.id === "string" ? event.id : void 0);
+  }
+  if (type.includes("tool")) {
+    const callId = pickString(data, ["toolCallId", "tool_use_id", "id"]) ?? (typeof event.id === "string" ? event.id : void 0) ?? `sandbox-tool-${++state.seq}`;
+    const toolName = pickString(data, ["name", "toolName", "tool"]) ?? "sandbox_tool";
+    if (type.includes("result")) {
+      return [
+        {
+          type: "tool_result",
+          toolName,
+          toolCallId: callId,
+          result: data.output ?? data.result ?? data.content ?? data
+        }
+      ];
+    }
+    return [
+      { type: "tool_call", toolName, toolCallId: callId, args: data.input ?? data.args ?? {} }
+    ];
+  }
+  return [];
+}
+function projectToolPart(part, state, eventId) {
+  const callId = pickString(part, ["callID", "callId", "toolCallId", "id"]) ?? eventId ?? `sandbox-tool-${++state.seq}`;
+  const toolName = pickString(part, ["tool", "toolName", "name"]) ?? "sandbox_tool";
+  const toolState = part.state && typeof part.state === "object" ? part.state : {};
+  const metadata = toolState.metadata && typeof toolState.metadata === "object" ? toolState.metadata : {};
+  const status = pickString(toolState, ["status"]) ?? "updated";
+  const previous = state.statusByCall.get(callId);
+  const settled = previous === "completed" || previous !== void 0 && TERMINAL_TOOL_FAILURE.test(previous);
+  if (settled) return [];
+  const out = [];
+  if (previous === void 0) {
+    out.push({
+      type: "tool_call",
+      toolName,
+      toolCallId: callId,
+      args: toolState.input ?? metadata.input ?? {}
+    });
+  }
+  state.statusByCall.set(callId, status);
+  if (status === "completed") {
+    out.push({
+      type: "tool_result",
+      toolName,
+      toolCallId: callId,
+      result: toolState.output ?? metadata.output ?? ""
+    });
+  } else if (TERMINAL_TOOL_FAILURE.test(status)) {
+    const message = pickString(toolState, ["error", "message"]) ?? pickString(metadata, ["error", "message"]) ?? `sandbox tool ended with status ${status}`;
+    const output = toolState.output ?? metadata.output;
+    out.push({
+      type: "tool_result",
+      toolName,
+      toolCallId: callId,
+      result: { error: message, status, ...output !== void 0 ? { output } : {} }
+    });
+  }
+  return out;
+}
+function pickString(data, keys) {
+  for (const key of keys) {
+    const value = data[key];
+    if (typeof value === "string" && value.length > 0) return value;
+  }
+  return void 0;
+}
 function mapSandboxEvent(event, opts = {}) {
   if (!event || typeof event !== "object") return void 0;
   const type = String(event.type ?? "");
@@ -87,6 +164,8 @@ function mapSandboxEvent(event, opts = {}) {
 export {
   extractLlmCallEvent,
   sumSandboxUsage,
+  createSandboxToolPartState,
+  mapSandboxToolEvent,
   mapSandboxEvent
 };
-//# sourceMappingURL=chunk-4J6RBI3K.js.map
+//# sourceMappingURL=chunk-QUXZBZCS.js.map

package/dist/chunk-QUXZBZCS.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/runtime/sandbox-events.ts"],"sourcesContent":["/**\n * Sandbox-event → runtime-event mapping.\n *\n * The sandbox SDK emits a polymorphic `SandboxEvent = { type, data, id? }`\n * whose `type` vocabulary is backend-determined (opencode, etc.) rather than\n * enumerated by the SDK. Two consumers project it:\n *   - the loop kernel's cost ledger (`extractLlmCallEvent`) — sums usage off\n *     every cost-bearing event, regardless of stream shape;\n *   - the `AgentRuntime.act` streaming contract (`mapSandboxEvent`) — projects\n *     incremental events to the `RuntimeStreamEvent` chat-UX vocabulary.\n *\n * Both live here so the empirically-observed `type` vocabulary has one home.\n */\n\nimport type { SandboxEvent } from '@tangle-network/sandbox'\nimport type { RuntimeStreamEvent } from '../types'\n\n/**\n * Extract a `RuntimeStreamEvent`-shaped `llm_call` from a sandbox event when\n * the event carries usage/cost data. Returns `undefined` for non-cost events\n * so the kernel can iterate the full stream without branching.\n *\n * Canonical cost-carrying types observed in the wild:\n *   - `llm_call` — `data: { model, tokensIn, tokensOut, costUsd, ... }`\n *   - `message.completed` / `result` — `data: { usage: { inputTokens,\n *      outputTokens, totalCostUsd? } }`\n *   - `cost.usage` / `usage` — same shape under a dedicated type\n *\n * Numeric coercion is strict: `Number.isFinite` gates every accumulator write\n * so a sentinel `NaN` from a misbehaving backend cannot poison the ledger.\n */\nexport function extractLlmCallEvent(\n  event: SandboxEvent,\n  agentRunName: string,\n): (RuntimeStreamEvent & { type: 'llm_call' }) | undefined {\n  if (!event || typeof event !== 'object') return undefined\n  const type = String(event.type ?? '')\n  const data =\n    event.data && typeof event.data === 'object'\n      ? (event.data as Record<string, unknown>)\n      : ({} as Record<string, unknown>)\n\n  if (type === 'llm_call' || type === 'cost.usage' || type === 'usage') {\n    return buildLlmCall(data, agentRunName)\n  }\n  if (type === 'message.completed' || type === 'result' || type === 'final') {\n    const usage = data.usage as Record<string, unknown> | undefined\n    if (!usage || typeof usage !== 'object') return undefined\n    return buildLlmCall({ ...usage, model: data.model ?? usage.model }, agentRunName)\n  }\n  // sandbox 0.4.0 terminal event: `data = { tokenUsage: { inputTokens, outputTokens,\n  // reasoningTokens, cacheReadInputTokens }, totalCostUsd }`. Usage lives under\n  // `tokenUsage` (not `usage`) and the cost is top-level — neither matched the\n  // branches above, so an in-process loopDispatch run reported {0,0} and the\n  // backend-integrity guard misread a real run as a stub. Reasoning tokens are\n  // billed output (reasoning models), so they fold into the output count.\n  if (type === 'done') {\n    const usage = data.tokenUsage as Record<string, unknown> | undefined\n    if (!usage || typeof usage !== 'object') return undefined\n    const out = pickFiniteNumber(usage, ['outputTokens', 'completion_tokens', 'tokensOut'])\n    const reasoning = pickFiniteNumber(usage, ['reasoningTokens'])\n    const mergedOut =\n      out !== undefined || reasoning !== undefined ? (out ?? 0) + (reasoning ?? 0) : undefined\n    return buildLlmCall(\n      {\n        inputTokens: usage.inputTokens,\n        outputTokens: mergedOut,\n        totalCostUsd: data.totalCostUsd,\n        model: data.model ?? usage.model,\n      },\n      agentRunName,\n    )\n  }\n  return undefined\n}\n\n/**\n * Sum the token usage + USD cost of a sandbox turn's events — the one honest way to meter an\n * `openSandboxRun` cell. Folds `extractLlmCallEvent` over the stream (which reads usage off EVERY backend\n * event shape), so a `runProfileMatrix` dispatch can report it to `ctx.cost`:\n *\n *     const turn = await run.start(prompt)\n *     const u = sumSandboxUsage(turn.events)\n *     if (u.input || u.output) ctx.cost.observeTokens({ input: u.input, output: u.output })\n *     if (u.costUsd) ctx.cost.observe(u.costUsd, 'sandbox-cell')\n *\n * Without this a cell reads `{tokens:0, cost:0}` and the backend-integrity guard correctly aborts the\n * matrix as a stub. `agentRunName` is the fallback model label for cost-only events (default `'agent'`).\n */\nexport function sumSandboxUsage(\n  events: readonly SandboxEvent[],\n  agentRunName = 'agent',\n): { input: number; output: number; costUsd: number } {\n  let input = 0\n  let output = 0\n  let costUsd = 0\n  for (const ev of events) {\n    const call = extractLlmCallEvent(ev, agentRunName)\n    if (!call) continue\n    input += call.tokensIn ?? 0\n    output += call.tokensOut ?? 0\n    costUsd += call.costUsd ?? 0\n  }\n  return { input, output, costUsd }\n}\n\nfunction buildLlmCall(\n  data: Record<string, unknown>,\n  agentRunName: string,\n): (RuntimeStreamEvent & { type: 'llm_call' }) | undefined {\n  const tokensIn = pickFiniteNumber(data, ['tokensIn', 'inputTokens', 'prompt_tokens'])\n  const tokensOut = pickFiniteNumber(data, ['tokensOut', 'outputTokens', 'completion_tokens'])\n  const costUsd = pickFiniteNumber(data, ['costUsd', 'totalCostUsd', 'cost_usd', 'cost'])\n  if (tokensIn === undefined && tokensOut === undefined && costUsd === undefined) {\n    return undefined\n  }\n  const model = typeof data.model === 'string' && data.model.length > 0 ? data.model : agentRunName\n  const event: RuntimeStreamEvent & { type: 'llm_call' } = {\n    type: 'llm_call',\n    model,\n  }\n  if (tokensIn !== undefined) event.tokensIn = tokensIn\n  if (tokensOut !== undefined) event.tokensOut = tokensOut\n  if (costUsd !== undefined) event.costUsd = costUsd\n  return event\n}\n\nfunction pickFiniteNumber(data: Record<string, unknown>, keys: string[]): number | undefined {\n  for (const key of keys) {\n    const value = data[key]\n    if (typeof value === 'number' && Number.isFinite(value)) return value\n  }\n  return undefined\n}\n\n/**\n * Cross-event state for {@link mapSandboxToolEvent}. Sandbox backends emit a\n * tool invocation as MANY `message.part.updated` frames on the same call id\n * (pending → running → completed), so faithful projection needs per-call\n * status memory: one `tool_call` on first sighting, at most one `tool_result`\n * on the terminal transition, nothing on intermediate re-frames. Create one\n * state per turn via {@link createSandboxToolPartState}.\n *\n * @experimental\n */\nexport interface SandboxToolPartState {\n  /** Last seen status per tool call id. A terminal status is sticky — later\n   *  frames on a settled call project to nothing. */\n  statusByCall: Map<string, string>\n  /** Sequence for synthesized call ids when an event carries none. */\n  seq: number\n}\n\n/** @experimental */\nexport function createSandboxToolPartState(): SandboxToolPartState {\n  return { statusByCall: new Map(), seq: 0 }\n}\n\n/** Terminal tool statuses that are failures (everything here settles the call). */\nconst TERMINAL_TOOL_FAILURE =\n  /^(error|errored|failed|failure|cancelled|canceled|timeout|timed_out)$/i\n\n/**\n * Project one `SandboxEvent` onto the `tool_call` / `tool_result` variants of\n * `RuntimeStreamEvent` — the tool-part projection `mapSandboxEvent`\n * deliberately does NOT perform. Opt-in and additive: `mapSandboxEvent`'s\n * default vocabulary (text/reasoning deltas + `llm_call`) is unchanged;\n * consumers that need the tool surface (chat UIs rendering tool activity)\n * compose this projector alongside it — `streamAgentTurn` does exactly that\n * under its `preserveToolParts` option.\n *\n * Handled shapes (observed on the opencode / claude-code sandbox backends):\n *   - `message.part.updated` with `part.type === 'tool'` — stateful: a\n *     `tool_call` on the call id's first frame (args from `state.input` or\n *     `state.metadata.input`), a `tool_result` when the status transitions to\n *     `completed` (result from `state.output` / `metadata.output`) or to a\n *     terminal failure (result is `{ error, status, output? }` — the error\n *     surfaced in-band, never dropped).\n *   - bare `tool*` event types (`tool.call`, `tool_result`, …) — stateless:\n *     `*result*` types project to `tool_result`, the rest to `tool_call`.\n *\n * Returns `[]` for every non-tool event.\n *\n * @experimental\n */\nexport function mapSandboxToolEvent(\n  event: SandboxEvent,\n  state: SandboxToolPartState,\n): (RuntimeStreamEvent & { type: 'tool_call' | 'tool_result' })[] {\n  if (!event || typeof event !== 'object') return []\n  const type = String(event.type ?? '')\n  const data =\n    event.data && typeof event.data === 'object'\n      ? (event.data as Record<string, unknown>)\n      : ({} as Record<string, unknown>)\n\n  if (type === 'message.part.updated') {\n    const part =\n      data.part && typeof data.part === 'object' ? (data.part as Record<string, unknown>) : {}\n    if (String(part.type ?? '') !== 'tool') return []\n    return projectToolPart(part, state, typeof event.id === 'string' ? event.id : undefined)\n  }\n\n  if (type.includes('tool')) {\n    const callId =\n      pickString(data, ['toolCallId', 'tool_use_id', 'id']) ??\n      (typeof event.id === 'string' ? event.id : undefined) ??\n      `sandbox-tool-${++state.seq}`\n    const toolName = pickString(data, ['name', 'toolName', 'tool']) ?? 'sandbox_tool'\n    if (type.includes('result')) {\n      return [\n        {\n          type: 'tool_result',\n          toolName,\n          toolCallId: callId,\n          result: data.output ?? data.result ?? data.content ?? data,\n        },\n      ]\n    }\n    return [\n      { type: 'tool_call', toolName, toolCallId: callId, args: data.input ?? data.args ?? {} },\n    ]\n  }\n\n  return []\n}\n\nfunction projectToolPart(\n  part: Record<string, unknown>,\n  state: SandboxToolPartState,\n  eventId: string | undefined,\n): (RuntimeStreamEvent & { type: 'tool_call' | 'tool_result' })[] {\n  const callId =\n    pickString(part, ['callID', 'callId', 'toolCallId', 'id']) ??\n    eventId ??\n    `sandbox-tool-${++state.seq}`\n  const toolName = pickString(part, ['tool', 'toolName', 'name']) ?? 'sandbox_tool'\n  const toolState =\n    part.state && typeof part.state === 'object' ? (part.state as Record<string, unknown>) : {}\n  const metadata =\n    toolState.metadata && typeof toolState.metadata === 'object'\n      ? (toolState.metadata as Record<string, unknown>)\n      : {}\n  const status = pickString(toolState, ['status']) ?? 'updated'\n\n  const previous = state.statusByCall.get(callId)\n  const settled =\n    previous === 'completed' || (previous !== undefined && TERMINAL_TOOL_FAILURE.test(previous))\n  if (settled) return []\n\n  const out: (RuntimeStreamEvent & { type: 'tool_call' | 'tool_result' })[] = []\n  if (previous === undefined) {\n    out.push({\n      type: 'tool_call',\n      toolName,\n      toolCallId: callId,\n      args: toolState.input ?? metadata.input ?? {},\n    })\n  }\n  state.statusByCall.set(callId, status)\n\n  if (status === 'completed') {\n    out.push({\n      type: 'tool_result',\n      toolName,\n      toolCallId: callId,\n      result: toolState.output ?? metadata.output ?? '',\n    })\n  } else if (TERMINAL_TOOL_FAILURE.test(status)) {\n    const message =\n      pickString(toolState, ['error', 'message']) ??\n      pickString(metadata, ['error', 'message']) ??\n      `sandbox tool ended with status ${status}`\n    const output = toolState.output ?? metadata.output\n    out.push({\n      type: 'tool_result',\n      toolName,\n      toolCallId: callId,\n      result: { error: message, status, ...(output !== undefined ? { output } : {}) },\n    })\n  }\n  return out\n}\n\nfunction pickString(data: Record<string, unknown>, keys: string[]): string | undefined {\n  for (const key of keys) {\n    const value = data[key]\n    if (typeof value === 'string' && value.length > 0) return value\n  }\n  return undefined\n}\n\n/**\n * Project one `SandboxEvent` onto the `RuntimeStreamEvent` chat-UX vocabulary,\n * for runtimes that bridge a sandbox `streamPrompt` into the\n * `AgentRuntime.act` streaming contract. Returns `undefined` for events that\n * have no faithful projection — the raw stream is preserved separately for the\n * `OutputAdapter`, so an unmapped event never loses data.\n *\n * Mapped (the task-optional incremental variants — no synthesized task\n * lifecycle, no guessed tool-part shapes):\n *   - `message.part.updated` text part → `text_delta`\n *   - `message.part.updated` reasoning/thinking part → `reasoning_delta`\n *   - cost-bearing events → `llm_call` (shared with the ledger extractor)\n *\n * Tool parts are deliberately NOT mapped here (unchanged default) — compose\n * {@link mapSandboxToolEvent} alongside when a consumer needs them.\n *\n * The opencode backend emits incremental text as\n * `{ type: 'message.part.updated', data: { part: { type, text }, delta } }`;\n * `delta` is the increment, `part.text` the running accumulation.\n */\nexport function mapSandboxEvent(\n  event: SandboxEvent,\n  opts: { agentRunName?: string } = {},\n): RuntimeStreamEvent | undefined {\n  if (!event || typeof event !== 'object') return undefined\n  const type = String(event.type ?? '')\n  const data =\n    event.data && typeof event.data === 'object'\n      ? (event.data as Record<string, unknown>)\n      : ({} as Record<string, unknown>)\n\n  if (type === 'message.part.updated') {\n    const part =\n      data.part && typeof data.part === 'object' ? (data.part as Record<string, unknown>) : {}\n    const partType = String(part.type ?? '')\n    const delta = typeof data.delta === 'string' ? data.delta : undefined\n    const text = delta ?? (typeof part.text === 'string' ? part.text : undefined)\n    if (text === undefined) return undefined\n    if (partType === 'text') return { type: 'text_delta', text }\n    if (partType === 'reasoning' || partType === 'thinking')\n      return { type: 'reasoning_delta', text }\n    return undefined\n  }\n\n  return extractLlmCallEvent(event, opts.agentRunName ?? 'agent')\n}\n"],"mappings":";AA+BO,SAAS,oBACd,OACA,cACyD;AACzD,MAAI,CAAC,SAAS,OAAO,UAAU,SAAU,QAAO;AAChD,QAAM,OAAO,OAAO,MAAM,QAAQ,EAAE;AACpC,QAAM,OACJ,MAAM,QAAQ,OAAO,MAAM,SAAS,WAC/B,MAAM,OACN,CAAC;AAER,MAAI,SAAS,cAAc,SAAS,gBAAgB,SAAS,SAAS;AACpE,WAAO,aAAa,MAAM,YAAY;AAAA,EACxC;AACA,MAAI,SAAS,uBAAuB,SAAS,YAAY,SAAS,SAAS;AACzE,UAAM,QAAQ,KAAK;AACnB,QAAI,CAAC,SAAS,OAAO,UAAU,SAAU,QAAO;AAChD,WAAO,aAAa,EAAE,GAAG,OAAO,OAAO,KAAK,SAAS,MAAM,MAAM,GAAG,YAAY;AAAA,EAClF;AAOA,MAAI,SAAS,QAAQ;AACnB,UAAM,QAAQ,KAAK;AACnB,QAAI,CAAC,SAAS,OAAO,UAAU,SAAU,QAAO;AAChD,UAAM,MAAM,iBAAiB,OAAO,CAAC,gBAAgB,qBAAqB,WAAW,CAAC;AACtF,UAAM,YAAY,iBAAiB,OAAO,CAAC,iBAAiB,CAAC;AAC7D,UAAM,YACJ,QAAQ,UAAa,cAAc,UAAa,OAAO,MAAM,aAAa,KAAK;AACjF,WAAO;AAAA,MACL;AAAA,QACE,aAAa,MAAM;AAAA,QACnB,cAAc;AAAA,QACd,cAAc,KAAK;AAAA,QACnB,OAAO,KAAK,SAAS,MAAM;AAAA,MAC7B;AAAA,MACA;AAAA,IACF;AAAA,EACF;AACA,SAAO;AACT;AAeO,SAAS,gBACd,QACA,eAAe,SACqC;AACpD,MAAI,QAAQ;AACZ,MAAI,SAAS;AACb,MAAI,UAAU;AACd,aAAW,MAAM,QAAQ;AACvB,UAAM,OAAO,oBAAoB,IAAI,YAAY;AACjD,QAAI,CAAC,KAAM;AACX,aAAS,KAAK,YAAY;AAC1B,cAAU,KAAK,aAAa;AAC5B,eAAW,KAAK,WAAW;AAAA,EAC7B;AACA,SAAO,EAAE,OAAO,QAAQ,QAAQ;AAClC;AAEA,SAAS,aACP,MACA,cACyD;AACzD,QAAM,WAAW,iBAAiB,MAAM,CAAC,YAAY,eAAe,eAAe,CAAC;AACpF,QAAM,YAAY,iBAAiB,MAAM,CAAC,aAAa,gBAAgB,mBAAmB,CAAC;AAC3F,QAAM,UAAU,iBAAiB,MAAM,CAAC,WAAW,gBAAgB,YAAY,MAAM,CAAC;AACtF,MAAI,aAAa,UAAa,cAAc,UAAa,YAAY,QAAW;AAC9E,WAAO;AAAA,EACT;AACA,QAAM,QAAQ,OAAO,KAAK,UAAU,YAAY,KAAK,MAAM,SAAS,IAAI,KAAK,QAAQ;AACrF,QAAM,QAAmD;AAAA,IACvD,MAAM;AAAA,IACN;AAAA,EACF;AACA,MAAI,aAAa,OAAW,OAAM,WAAW;AAC7C,MAAI,cAAc,OAAW,OAAM,YAAY;AAC/C,MAAI,YAAY,OAAW,OAAM,UAAU;AAC3C,SAAO;AACT;AAEA,SAAS,iBAAiB,MAA+B,MAAoC;AAC3F,aAAW,OAAO,MAAM;AACtB,UAAM,QAAQ,KAAK,GAAG;AACtB,QAAI,OAAO,UAAU,YAAY,OAAO,SAAS,KAAK,EAAG,QAAO;AAAA,EAClE;AACA,SAAO;AACT;AAqBO,SAAS,6BAAmD;AACjE,SAAO,EAAE,cAAc,oBAAI,IAAI,GAAG,KAAK,EAAE;AAC3C;AAGA,IAAM,wBACJ;AAyBK,SAAS,oBACd,OACA,OACgE;AAChE,MAAI,CAAC,SAAS,OAAO,UAAU,SAAU,QAAO,CAAC;AACjD,QAAM,OAAO,OAAO,MAAM,QAAQ,EAAE;AACpC,QAAM,OACJ,MAAM,QAAQ,OAAO,MAAM,SAAS,WAC/B,MAAM,OACN,CAAC;AAER,MAAI,SAAS,wBAAwB;AACnC,UAAM,OACJ,KAAK,QAAQ,OAAO,KAAK,SAAS,WAAY,KAAK,OAAmC,CAAC;AACzF,QAAI,OAAO,KAAK,QAAQ,EAAE,MAAM,OAAQ,QAAO,CAAC;AAChD,WAAO,gBAAgB,MAAM,OAAO,OAAO,MAAM,OAAO,WAAW,MAAM,KAAK,MAAS;AAAA,EACzF;AAEA,MAAI,KAAK,SAAS,MAAM,GAAG;AACzB,UAAM,SACJ,WAAW,MAAM,CAAC,cAAc,eAAe,IAAI,CAAC,MACnD,OAAO,MAAM,OAAO,WAAW,MAAM,KAAK,WAC3C,gBAAgB,EAAE,MAAM,GAAG;AAC7B,UAAM,WAAW,WAAW,MAAM,CAAC,QAAQ,YAAY,MAAM,CAAC,KAAK;AACnE,QAAI,KAAK,SAAS,QAAQ,GAAG;AAC3B,aAAO;AAAA,QACL;AAAA,UACE,MAAM;AAAA,UACN;AAAA,UACA,YAAY;AAAA,UACZ,QAAQ,KAAK,UAAU,KAAK,UAAU,KAAK,WAAW;AAAA,QACxD;AAAA,MACF;AAAA,IACF;AACA,WAAO;AAAA,MACL,EAAE,MAAM,aAAa,UAAU,YAAY,QAAQ,MAAM,KAAK,SAAS,KAAK,QAAQ,CAAC,EAAE;AAAA,IACzF;AAAA,EACF;AAEA,SAAO,CAAC;AACV;AAEA,SAAS,gBACP,MACA,OACA,SACgE;AAChE,QAAM,SACJ,WAAW,MAAM,CAAC,UAAU,UAAU,cAAc,IAAI,CAAC,KACzD,WACA,gBAAgB,EAAE,MAAM,GAAG;AAC7B,QAAM,WAAW,WAAW,MAAM,CAAC,QAAQ,YAAY,MAAM,CAAC,KAAK;AACnE,QAAM,YACJ,KAAK,SAAS,OAAO,KAAK,UAAU,WAAY,KAAK,QAAoC,CAAC;AAC5F,QAAM,WACJ,UAAU,YAAY,OAAO,UAAU,aAAa,WAC/C,UAAU,WACX,CAAC;AACP,QAAM,SAAS,WAAW,WAAW,CAAC,QAAQ,CAAC,KAAK;AAEpD,QAAM,WAAW,MAAM,aAAa,IAAI,MAAM;AAC9C,QAAM,UACJ,aAAa,eAAgB,aAAa,UAAa,sBAAsB,KAAK,QAAQ;AAC5F,MAAI,QAAS,QAAO,CAAC;AAErB,QAAM,MAAsE,CAAC;AAC7E,MAAI,aAAa,QAAW;AAC1B,QAAI,KAAK;AAAA,MACP,MAAM;AAAA,MACN;AAAA,MACA,YAAY;AAAA,MACZ,MAAM,UAAU,SAAS,SAAS,SAAS,CAAC;AAAA,IAC9C,CAAC;AAAA,EACH;AACA,QAAM,aAAa,IAAI,QAAQ,MAAM;AAErC,MAAI,WAAW,aAAa;AAC1B,QAAI,KAAK;AAAA,MACP,MAAM;AAAA,MACN;AAAA,MACA,YAAY;AAAA,MACZ,QAAQ,UAAU,UAAU,SAAS,UAAU;AAAA,IACjD,CAAC;AAAA,EACH,WAAW,sBAAsB,KAAK,MAAM,GAAG;AAC7C,UAAM,UACJ,WAAW,WAAW,CAAC,SAAS,SAAS,CAAC,KAC1C,WAAW,UAAU,CAAC,SAAS,SAAS,CAAC,KACzC,kCAAkC,MAAM;AAC1C,UAAM,SAAS,UAAU,UAAU,SAAS;AAC5C,QAAI,KAAK;AAAA,MACP,MAAM;AAAA,MACN;AAAA,MACA,YAAY;AAAA,MACZ,QAAQ,EAAE,OAAO,SAAS,QAAQ,GAAI,WAAW,SAAY,EAAE,OAAO,IAAI,CAAC,EAAG;AAAA,IAChF,CAAC;AAAA,EACH;AACA,SAAO;AACT;AAEA,SAAS,WAAW,MAA+B,MAAoC;AACrF,aAAW,OAAO,MAAM;AACtB,UAAM,QAAQ,KAAK,GAAG;AACtB,QAAI,OAAO,UAAU,YAAY,MAAM,SAAS,EAAG,QAAO;AAAA,EAC5D;AACA,SAAO;AACT;AAsBO,SAAS,gBACd,OACA,OAAkC,CAAC,GACH;AAChC,MAAI,CAAC,SAAS,OAAO,UAAU,SAAU,QAAO;AAChD,QAAM,OAAO,OAAO,MAAM,QAAQ,EAAE;AACpC,QAAM,OACJ,MAAM,QAAQ,OAAO,MAAM,SAAS,WAC/B,MAAM,OACN,CAAC;AAER,MAAI,SAAS,wBAAwB;AACnC,UAAM,OACJ,KAAK,QAAQ,OAAO,KAAK,SAAS,WAAY,KAAK,OAAmC,CAAC;AACzF,UAAM,WAAW,OAAO,KAAK,QAAQ,EAAE;AACvC,UAAM,QAAQ,OAAO,KAAK,UAAU,WAAW,KAAK,QAAQ;AAC5D,UAAM,OAAO,UAAU,OAAO,KAAK,SAAS,WAAW,KAAK,OAAO;AACnE,QAAI,SAAS,OAAW,QAAO;AAC/B,QAAI,aAAa,OAAQ,QAAO,EAAE,MAAM,cAAc,KAAK;AAC3D,QAAI,aAAa,eAAe,aAAa;AAC3C,aAAO,EAAE,MAAM,mBAAmB,KAAK;AACzC,WAAO;AAAA,EACT;AAEA,SAAO,oBAAoB,OAAO,KAAK,gBAAgB,OAAO;AAChE;","names":[]}

package/dist/{chunk-DLAEEF26.js → chunk-VKVNDNG4.js}

@@ -15,7 +15,7 @@ import {
   settledToIteration,
   supervise,
   withDriverExecutor
-} from "./chunk-XN5TIJGP.js";
+} from "./chunk-6XCW3M7W.js";
 import {
   addTokenUsage,
   isAbortError,
@@ -25,8 +25,10 @@ import {
   zeroTokenUsage
 } from "./chunk-BZF3KQ6G.js";
 import {
-  mapSandboxEvent
-} from "./chunk-4J6RBI3K.js";
+  createSandboxToolPartState,
+  mapSandboxEvent,
+  mapSandboxToolEvent
+} from "./chunk-QUXZBZCS.js";
 import {
   AnalystError,
   BackendTransportError,
@@ -977,49 +979,68 @@ function defineLeaderboard(spec) {
       return p;
     };
     let shotNonce = 0;
-    const dispatch = spec.dispatch ?? loopDispatch({
-      sandboxClient,
-      toLoopOptions: (scenario, profile) => {
-        const axis = harnessAxisOf(profile);
-        const modelId = bareModel(axis?.model ?? models[0] ?? "");
-        return {
-          // naiveDriver = the no-signal retry floor: re-run the same case as
-          // an independent attempt until one scores (>0) or the shot cap.
-          driver: naiveDriver({
-            continuation: "",
-            applyContinuation: (task) => task,
-            maxIterations: shots
-          }),
-          agentRun: {
-            profile,
-            taskToPrompt: (s) => `${promptOf(s)}
+    const dispatch = spec.dispatch ?? ((profile, scenario, dispatchCtx) => {
+      const cellDispatch = loopDispatch({
+        sandboxClient,
+        toLoopOptions: (cellScenario, cellProfile) => {
+          const axis = harnessAxisOf(cellProfile);
+          const modelId = bareModel(axis?.model ?? models[0] ?? "");
+          return {
+            // naiveDriver = the no-signal retry floor: re-run the same case as
+            // an independent attempt until one scores (>0) or the shot cap.
+            driver: naiveDriver({
+              continuation: "",
+              applyContinuation: (task) => task,
+              maxIterations: shots
+            }),
+            agentRun: {
+              profile: cellProfile,
+              taskToPrompt: (s) => `${promptOf(s)}
 
 <!-- independent-attempt:${shotNonce++} -->`,
-            ...axis ? {
-              sandboxOverrides: {
-                backend: {
-                  type: axis.harness,
-                  model: { ...spec.modelBackend, model: modelId }
+              ...axis ? {
+                sandboxOverrides: {
+                  backend: {
+                    type: axis.harness,
+                    model: { ...spec.modelBackend, model: modelId }
+                  }
                 }
+              } : {}
+            },
+            output: {
+              parse: (events) => spec.parseOutput ? spec.parseOutput(events, cellScenario.case) : (
+                // The default decode produces string — the TArtifact
+                // default. A structured-TArtifact spec supplies parseOutput
+                // (documented on the field), so this cast never lies.
+                collectAgentResponseText(events) ?? ""
+              )
+            },
+            validator: {
+              validate: async (output) => {
+                const s = normalizeScore(spec.score(output, cellScenario.case));
+                return { valid: s.composite > 0, score: s.composite };
               }
-            } : {}
-          },
-          output: {
-            parse: (events) => {
-              spec.onCellEvents?.(events, scenario.case);
-              return spec.parseOutput ? spec.parseOutput(events, scenario.case) : collectAgentResponseText(events) ?? "";
-            }
-          },
-          validator: {
-            validate: async (output) => {
-              const s = normalizeScore(spec.score(output, scenario.case));
-              return { valid: s.composite > 0, score: s.composite };
+            },
+            task: cellScenario,
+            maxIterations: shots
+          };
+        },
+        toArtifact: (result) => {
+          for (const iter of result.iterations) {
+            spec.onCellEvents?.(iter.events, scenario.case, {
+              index: iter.index,
+              ...iter.error ? { error: iter.error.message } : {},
+              ...iter.verdict ? { verdict: { score: iter.verdict.score } } : {}
+            });
+            if (spec.resolveModel) {
+              const served = spec.resolveModel(iter.events);
+              if (served !== void 0) dispatchCtx.cost.observeModel?.(served);
             }
-          },
-          task: scenario,
-          maxIterations: shots
-        };
-      }
+          }
+          return result.winner?.output;
+        }
+      });
+      return cellDispatch(profile, scenario, dispatchCtx);
     });
     await spec.setup?.(ctx);
     try {
@@ -1306,7 +1327,7 @@ function isAsyncIterable2(v) {
   return typeof v === "object" && v !== null && Symbol.asyncIterator in v;
 }
 function inProcessSandboxClient(options) {
-  const { onPrompt, workdir: workdirPrefix, id: idOption } = options;
+  const { onPrompt, onTask, workdir: workdirPrefix, id: idOption } = options;
   let seq = 0;
   return {
     async create(_options) {
@@ -1345,19 +1366,34 @@ function inProcessSandboxClient(options) {
           }
         }
       } : {};
+      async function* drive(behavior, mode, message, opts) {
+        const prompt = typeof message === "string" ? message : JSON.stringify(message);
+        const { signal: optSignal, ...rest } = opts ?? {};
+        const signal = optSignal ?? new AbortController().signal;
+        const ctx = {
+          round,
+          workdir: boxWorkdir,
+          signal,
+          mode,
+          ...Object.keys(rest).length > 0 ? { options: rest } : {}
+        };
+        round += 1;
+        const produced = await behavior(prompt, ctx);
+        if (isAsyncIterable2(produced)) {
+          for await (const ev of produced) yield ev;
+        } else {
+          for (const ev of produced) yield ev;
+        }
+      }
       const box = {
         id,
-        async *streamPrompt(message, opts) {
-          const prompt = typeof message === "string" ? message : JSON.stringify(message);
-          const signal = opts?.signal ?? new AbortController().signal;
-          const ctx = { round, workdir: boxWorkdir, signal };
-          round += 1;
-          const produced = await onPrompt(prompt, ctx);
-          if (isAsyncIterable2(produced)) {
-            for await (const ev of produced) yield ev;
-          } else {
-            for (const ev of produced) yield ev;
-          }
+        streamPrompt(message, opts) {
+          return drive(onPrompt, "prompt", message, opts);
+        },
+        // Task-mode verb (`streamAgentTurn`'s `box-task` backend). Drives
+        // `onTask` when configured; otherwise the box's one behavior callback.
+        streamTask(message, opts) {
+          return drive(onTask ?? onPrompt, "task", message, opts);
         },
         ...fsMembers,
         async delete() {
@@ -4301,6 +4337,12 @@ function createOpenAICompatibleBackend(options) {
       if (options.responseFormat !== void 0) {
         bodyPayload.response_format = options.responseFormat;
       }
+      if (options.temperature !== void 0) {
+        bodyPayload.temperature = options.temperature;
+      }
+      if (options.maxTokens !== void 0) {
+        bodyPayload.max_tokens = options.maxTokens;
+      }
       const requestBody = JSON.stringify(bodyPayload);
       let response;
       let lastStatus = 0;
@@ -4781,11 +4823,17 @@ async function* streamAgentTurn(backend, prompt, opts = {}) {
     session = await startTurnSession(backend, task, prompt, deadline.signal, label);
     yield { type: "backend_start", task, session, backend: label, timestamp: nowIso() };
     const inner = backend.kind === "chat" ? driveChatTurn(backend.backend, task, session, prompt, deadline.signal, acc) : driveBoxTurn(
-      backend.kind === "box" ? backend.box : await inlineSandboxClient(backend.factory).create(),
+      backend.kind === "executor" ? await inlineSandboxClient(backend.factory).create() : backend.box,
       prompt,
       deadline.signal,
       backend.agentRunName ?? "agent",
-      acc
+      acc,
+      {
+        mode: backend.kind === "box-task" ? "task" : "prompt",
+        ...backend.kind !== "executor" && backend.options ? { options: backend.options } : {},
+        preserveToolParts: opts.preserveToolParts === true,
+        ...opts.onRawEvent ? { onRawEvent: opts.onRawEvent } : {}
+      }
     );
     for await (const event of inner) {
       yield event;
@@ -4849,10 +4897,17 @@ async function startTurnSession(backend, task, prompt, signal, label) {
   }
   return newRuntimeSession(label);
 }
-async function* driveBoxTurn(box, prompt, signal, agentRunName, acc) {
-  for await (const event of box.streamPrompt(prompt, { signal })) {
+async function* driveBoxTurn(box, prompt, signal, agentRunName, acc, cfg) {
+  const callOptions = { ...cfg.options ?? {}, signal };
+  const stream = cfg.mode === "task" ? box.streamTask(prompt, callOptions) : box.streamPrompt(prompt, callOptions);
+  const toolParts = cfg.preserveToolParts ? createSandboxToolPartState() : void 0;
+  for await (const event of stream) {
+    if (cfg.onRawEvent) await cfg.onRawEvent(event);
     const terminalText = terminalTextFromSandboxEvent(event);
     if (terminalText !== void 0) acc.terminalText = terminalText;
+    if (toolParts) {
+      for (const toolEvent of mapSandboxToolEvent(event, toolParts)) yield toolEvent;
+    }
     const mapped = mapSandboxEvent(event, { agentRunName });
     if (!mapped) continue;
     foldEvent(mapped, acc, agentRunName);
@@ -5851,4 +5906,4 @@ export {
   computeFindingId,
   makeFinding2 as makeFinding
 };
-//# sourceMappingURL=chunk-DLAEEF26.js.map
+//# sourceMappingURL=chunk-VKVNDNG4.js.map