@psnext/slingcli 2.5.20260722-2 → 2.5.20260723-1

package/node_modules/@earendil-works/pi-agent-core/dist/agent-loop.js

@@ -2,7 +2,8 @@
  * Agent loop that works with AgentMessage throughout.
  * Transforms to Message[] only at the LLM call boundary.
  */
-import { EventStream, streamSimple, validateToolArguments, } from "@earendil-works/pi-ai/compat";
+import { EventStream, validateToolArguments, } from "@earendil-works/pi-ai";
+import { getDefaultStreamFn } from "./stream-fn.js";
 /**
  * Start an agent loop with a new prompt message.
  * The prompt is added to the context and events are emitted for it.
@@ -51,7 +52,7 @@ export async function runAgentLoop(prompts, context, config, emit, signal, strea
         await emit({ type: "message_start", message: prompt });
         await emit({ type: "message_end", message: prompt });
     }
-    await runLoop(currentContext, newMessages, config, signal, emit, streamFn);
+    await runLoop(currentContext, newMessages, config, signal, emit, streamFn ?? getDefaultStreamFn());
     return newMessages;
 }
 export async function runAgentLoopContinue(context, config, emit, signal, streamFn) {
@@ -65,7 +66,7 @@ export async function runAgentLoopContinue(context, config, emit, signal, stream
     const currentContext = { ...context };
     await emit({ type: "agent_start" });
     await emit({ type: "turn_start" });
-    await runLoop(currentContext, newMessages, config, signal, emit, streamFn);
+    await runLoop(currentContext, newMessages, config, signal, emit, streamFn ?? getDefaultStreamFn());
     return newMessages;
 }
 function createAgentStream() {
@@ -74,7 +75,7 @@ function createAgentStream() {
 /**
  * Main loop logic shared by agentLoop and agentLoopContinue.
  */
-async function runLoop(initialContext, newMessages, initialConfig, signal, emit, streamFn) {
+async function runLoop(initialContext, newMessages, initialConfig, signal, emit, streamFunction) {
     let currentContext = initialContext;
     let config = initialConfig;
     let firstTurn = true;
@@ -102,7 +103,7 @@ async function runLoop(initialContext, newMessages, initialConfig, signal, emit,
                 pendingMessages = [];
             }
             // Stream assistant response
-            const message = await streamAssistantResponse(currentContext, config, signal, emit, streamFn);
+            const message = await streamAssistantResponse(currentContext, config, signal, emit, streamFunction);
             newMessages.push(message);
             if (message.stopReason === "error" || message.stopReason === "aborted") {
                 await emit({ type: "turn_end", message, toolResults: [] });
@@ -174,7 +175,7 @@ async function runLoop(initialContext, newMessages, initialConfig, signal, emit,
  * Stream an assistant response from the LLM.
  * This is where AgentMessage[] gets transformed to Message[] for the LLM.
  */
-async function streamAssistantResponse(context, config, signal, emit, streamFn) {
+async function streamAssistantResponse(context, config, signal, emit, streamFunction) {
     // Apply context transform if configured (AgentMessage[] → AgentMessage[])
     let messages = context.messages;
     if (config.transformContext) {
@@ -188,7 +189,6 @@ async function streamAssistantResponse(context, config, signal, emit, streamFn)
         messages: llmMessages,
         tools: context.tools,
     };
-    const streamFunction = streamFn || streamSimple;
     // Resolve API key (important for expiring tokens)
     const resolvedApiKey = (config.getApiKey ? await config.getApiKey(config.model.provider) : undefined) || config.apiKey;
     const response = await streamFunction(config.model, llmContext, {
@@ -495,6 +495,7 @@ async function finalizeExecutedToolCall(currentContext, assistantMessage, prepar
                     ...result,
                     content: afterResult.content ?? result.content,
                     details: afterResult.details ?? result.details,
+                    usage: afterResult.usage ?? result.usage,
                     terminate: afterResult.terminate ?? result.terminate,
                 };
                 isError = afterResult.isError ?? isError;
@@ -535,6 +536,7 @@ function createToolResultMessage(finalized) {
         // so the null never enters session history or provider payloads.
         content: finalized.result.content ?? [],
         details: finalized.result.details,
+        usage: finalized.result.usage,
         ...(finalized.result.addedToolNames?.length ? { addedToolNames: finalized.result.addedToolNames } : {}),
         isError: finalized.isError,
         timestamp: Date.now(),

package/node_modules/@earendil-works/pi-agent-core/dist/agent.js

@@ -1,5 +1,5 @@
-import { streamSimple, } from "@earendil-works/pi-ai/compat";
 import { runAgentLoop, runAgentLoopContinue } from "./agent-loop.js";
+import { getDefaultStreamFn } from "./stream-fn.js";
 function defaultConvertToLlm(messages) {
     return messages.filter((message) => message.role === "user" || message.role === "assistant" || message.role === "toolResult");
 }
@@ -90,7 +90,7 @@ export class Agent {
     followUpQueue;
     convertToLlm;
     transformContext;
-    streamFn;
+    streamFunction;
     getApiKey;
     onPayload;
     onResponse;
@@ -109,25 +109,27 @@ export class Agent {
     maxRetryDelayMs;
     /** Tool execution strategy for assistant messages that contain multiple tool calls. */
     toolExecution;
-    constructor(options = {}) {
-        this._state = createMutableAgentState(options.initialState);
-        this.convertToLlm = options.convertToLlm ?? defaultConvertToLlm;
-        this.transformContext = options.transformContext;
-        this.streamFn = options.streamFn ?? streamSimple;
-        this.getApiKey = options.getApiKey;
-        this.onPayload = options.onPayload;
-        this.onResponse = options.onResponse;
-        this.beforeToolCall = options.beforeToolCall;
-        this.afterToolCall = options.afterToolCall;
-        this.prepareNextTurn = options.prepareNextTurn;
-        this.prepareNextTurnWithContext = options.prepareNextTurnWithContext;
-        this.steeringQueue = new PendingMessageQueue(options.steeringMode ?? "one-at-a-time");
-        this.followUpQueue = new PendingMessageQueue(options.followUpMode ?? "one-at-a-time");
-        this.sessionId = options.sessionId;
-        this.thinkingBudgets = options.thinkingBudgets;
-        this.transport = options.transport ?? "auto";
-        this.maxRetryDelayMs = options.maxRetryDelayMs;
-        this.toolExecution = options.toolExecution ?? "parallel";
+    constructor(options) {
+        // Older compiled consumers may omit options or streamFn even though the current API requires them.
+        const runtimeOptions = options ?? {};
+        this._state = createMutableAgentState(runtimeOptions.initialState);
+        this.convertToLlm = runtimeOptions.convertToLlm ?? defaultConvertToLlm;
+        this.transformContext = runtimeOptions.transformContext;
+        this.streamFunction = runtimeOptions.streamFn ?? getDefaultStreamFn();
+        this.getApiKey = runtimeOptions.getApiKey;
+        this.onPayload = runtimeOptions.onPayload;
+        this.onResponse = runtimeOptions.onResponse;
+        this.beforeToolCall = runtimeOptions.beforeToolCall;
+        this.afterToolCall = runtimeOptions.afterToolCall;
+        this.prepareNextTurn = runtimeOptions.prepareNextTurn;
+        this.prepareNextTurnWithContext = runtimeOptions.prepareNextTurnWithContext;
+        this.steeringQueue = new PendingMessageQueue(runtimeOptions.steeringMode ?? "one-at-a-time");
+        this.followUpQueue = new PendingMessageQueue(runtimeOptions.followUpMode ?? "one-at-a-time");
+        this.sessionId = runtimeOptions.sessionId;
+        this.thinkingBudgets = runtimeOptions.thinkingBudgets;
+        this.transport = runtimeOptions.transport ?? "auto";
+        this.maxRetryDelayMs = runtimeOptions.maxRetryDelayMs;
+        this.toolExecution = runtimeOptions.toolExecution ?? "parallel";
     }
     /**
      * Subscribe to agent lifecycle events.
@@ -262,12 +264,12 @@ export class Agent {
     }
     async runPromptMessages(messages, options = {}) {
         await this.runWithLifecycle(async (signal) => {
-            await runAgentLoop(messages, this.createContextSnapshot(), this.createLoopConfig(options), (event) => this.processEvents(event), signal, this.streamFn);
+            await runAgentLoop(messages, this.createContextSnapshot(), this.createLoopConfig(options), (event) => this.processEvents(event), signal, this.streamFunction);
         });
     }
     async runContinuation() {
         await this.runWithLifecycle(async (signal) => {
-            await runAgentLoopContinue(this.createContextSnapshot(), this.createLoopConfig(), (event) => this.processEvents(event), signal, this.streamFn);
+            await runAgentLoopContinue(this.createContextSnapshot(), this.createLoopConfig(), (event) => this.processEvents(event), signal, this.streamFunction);
         });
     }
     createContextSnapshot() {

package/node_modules/@earendil-works/pi-agent-core/dist/harness/agent-harness.js

@@ -1,3 +1,4 @@
+import { contentText, } from "@earendil-works/pi-ai";
 import { runAgentLoop } from "../agent-loop.js";
 import { collectEntriesForBranchSummary, generateBranchSummary } from "./compaction/branch-summarization.js";
 import { compact, DEFAULT_COMPACTION_SETTINGS, prepareCompaction } from "./compaction/compaction.js";
@@ -122,6 +123,7 @@ export class AgentHarness {
     thinkingLevel;
     systemPrompt;
     streamOptions;
+    retry;
     resources;
     tools = new Map();
     activeToolNames;
@@ -137,6 +139,7 @@ export class AgentHarness {
         this.models = options.models;
         this.resources = options.resources ?? {};
         this.streamOptions = cloneStreamOptions(options.streamOptions);
+        this.retry = options.retry;
         this.systemPrompt = options.systemPrompt;
         this.validateUniqueNames((options.tools ?? []).map((tool) => tool.name), "Duplicate tool name(s)");
         for (const tool of options.tools ?? []) {
@@ -193,6 +196,13 @@ export class AgentHarness {
         }
         return lastResult;
     }
+    retryCallbacks(operation) {
+        return {
+            onRetryScheduled: (attempt, maxAttempts, delayMs, errorMessage) => this.emitOwn({ type: "retry_scheduled", operation, attempt, maxAttempts, delayMs, errorMessage }),
+            onRetryAttemptStart: () => this.emitOwn({ type: "retry_attempt_start", operation }),
+            onRetryFinished: () => this.emitOwn({ type: "retry_finished", operation }),
+        };
+    }
     async emitBeforeProviderRequest(model, sessionId, streamOptions) {
         const handlers = this.getHandlers("before_provider_request");
         let current = cloneStreamOptions(streamOptions);
@@ -358,9 +368,16 @@ export class AgentHarness {
                     content: result.content,
                     details: result.details,
                     isError,
+                    usage: result.usage,
                 });
                 return patch
-                    ? { content: patch.content, details: patch.details, isError: patch.isError, terminate: patch.terminate }
+                    ? {
+                        content: patch.content,
+                        details: patch.details,
+                        isError: patch.isError,
+                        usage: patch.usage,
+                        terminate: patch.terminate,
+                    }
                     : undefined;
             },
             prepareNextTurn: async () => {
@@ -634,11 +651,11 @@ export class AgentHarness {
             const provided = hookResult?.compaction;
             const compactResult = provided
                 ? { ok: true, value: provided }
-                : await compact(preparation, this.models, model, customInstructions, undefined, this.thinkingLevel);
+                : await compact(preparation, this.models, model, customInstructions, undefined, this.thinkingLevel, this.retry, this.retryCallbacks("compaction"));
             if (!compactResult.ok)
                 throw compactResult.error;
             const result = compactResult.value;
-            const entryId = await this.session.appendCompaction(result.summary, result.firstKeptEntryId, result.tokensBefore, result.details, provided !== undefined);
+            const entryId = await this.session.appendCompaction(result.summary, result.firstKeptEntryId, result.tokensBefore, result.details, provided !== undefined, result.usage, result.retainedTail);
             const entry = await this.session.getEntry(entryId);
             if (entry?.type === "compaction") {
                 await this.emitOwn({ type: "session_compact", compactionEntry: entry, fromHook: provided !== undefined });
@@ -681,6 +698,7 @@ export class AgentHarness {
             let summaryEntry;
             let summaryText = hookResult?.summary?.summary;
             let summaryDetails = hookResult?.summary?.details;
+            let summaryUsage = hookResult?.summary?.usage;
             if (!summaryText && options?.summarize && entries.length > 0) {
                 const model = this.model;
                 if (!model)
@@ -691,6 +709,8 @@ export class AgentHarness {
                     signal: new AbortController().signal,
                     customInstructions: hookResult?.customInstructions ?? options?.customInstructions,
                     replaceInstructions: hookResult?.replaceInstructions ?? options?.replaceInstructions,
+                    retry: this.retry,
+                    callbacks: this.retryCallbacks("branch_summary"),
                 });
                 if (!branchSummary.ok) {
                     if (branchSummary.error.code === "aborted")
@@ -698,6 +718,7 @@ export class AgentHarness {
                     throw new AgentHarnessError("branch_summary", branchSummary.error.message, branchSummary.error);
                 }
                 summaryText = branchSummary.value.summary;
+                summaryUsage = branchSummary.value.usage;
                 summaryDetails = {
                     readFiles: branchSummary.value.readFiles,
                     modifiedFiles: branchSummary.value.modifiedFiles,
@@ -707,30 +728,22 @@ export class AgentHarness {
             let newLeafId;
             if (targetEntry.type === "message" && targetEntry.message.role === "user") {
                 newLeafId = targetEntry.parentId;
-                const content = targetEntry.message.content;
-                editorText =
-                    typeof content === "string"
-                        ? content
-                        : content
-                            .filter((c) => c.type === "text")
-                            .map((c) => c.text)
-                            .join("");
+                editorText = contentText(targetEntry.message.content, "");
             }
             else if (targetEntry.type === "custom_message") {
                 newLeafId = targetEntry.parentId;
-                editorText =
-                    typeof targetEntry.content === "string"
-                        ? targetEntry.content
-                        : targetEntry.content
-                            .filter((c) => c.type === "text")
-                            .map((c) => c.text)
-                            .join("");
+                editorText = contentText(targetEntry.content, "");
             }
             else {
                 newLeafId = targetId;
             }
             const summaryId = await this.session.moveTo(newLeafId, summaryText
-                ? { summary: summaryText, details: summaryDetails, fromHook: hookResult?.summary !== undefined }
+                ? {
+                    summary: summaryText,
+                    details: summaryDetails,
+                    usage: summaryUsage,
+                    fromHook: hookResult?.summary !== undefined,
+                }
                 : undefined);
             if (summaryId) {
                 const entry = await this.session.getEntry(summaryId);

package/node_modules/@earendil-works/pi-agent-core/dist/harness/compaction/branch-summarization.js

@@ -1,6 +1,7 @@
+import { contentText } from "@earendil-works/pi-ai";
 import { convertToLlm, createBranchSummaryMessage, createCompactionSummaryMessage, createCustomMessage, } from "../messages.js";
 import { BranchSummaryError, err, ok, SessionError } from "../types.js";
-import { estimateTokens, SUMMARIZATION_SYSTEM_PROMPT } from "./compaction.js";
+import { completeSimpleWithRetries, estimateTokens, SUMMARIZATION_SYSTEM_PROMPT } from "./compaction.js";
 import { computeFileLists, createFileOps, extractFileOpsFromMessage, formatFileOperations, serializeConversation, } from "./utils.js";
 /** Collect entries that should be summarized before navigating to a different session tree entry. */
 export async function collectEntriesForBranchSummary(session, oldLeafId, targetId) {
@@ -124,7 +125,7 @@ Use this EXACT format:
 Keep each section concise. Preserve exact file paths, function names, and error messages.`;
 /** Generate a summary for abandoned branch entries. */
 export async function generateBranchSummary(entries, options) {
-    const { models, model, signal, customInstructions, replaceInstructions, reserveTokens = 16384 } = options;
+    const { models, model, signal, customInstructions, replaceInstructions, reserveTokens = 16384, retry, callbacks, } = options;
     const contextWindow = model.contextWindow || 128000;
     const tokenBudget = contextWindow - reserveTokens;
     const { messages, fileOps } = prepareBranchEntries(entries, tokenBudget);
@@ -151,22 +152,20 @@ export async function generateBranchSummary(entries, options) {
             timestamp: Date.now(),
         },
     ];
-    const response = await models.completeSimple(model, { systemPrompt: SUMMARIZATION_SYSTEM_PROMPT, messages: summarizationMessages }, { signal, maxTokens: 2048 });
+    const response = await completeSimpleWithRetries(models, model, { systemPrompt: SUMMARIZATION_SYSTEM_PROMPT, messages: summarizationMessages }, { signal, maxTokens: 2048 }, retry, callbacks);
     if (response.stopReason === "aborted") {
         return err(new BranchSummaryError("aborted", response.errorMessage || "Branch summary aborted"));
     }
     if (response.stopReason === "error") {
         return err(new BranchSummaryError("summarization_failed", `Branch summary failed: ${response.errorMessage || "Unknown error"}`));
     }
-    let summary = response.content
-        .filter((c) => c.type === "text")
-        .map((c) => c.text)
-        .join("\n");
+    let summary = contentText(response.content);
     summary = BRANCH_SUMMARY_PREAMBLE + summary;
     const { readFiles, modifiedFiles } = computeFileLists(fileOps);
     summary += formatFileOperations(readFiles, modifiedFiles);
     return ok({
         summary: summary || "No summary generated",
+        usage: response.usage,
         readFiles,
         modifiedFiles,
     });

package/node_modules/@earendil-works/pi-agent-core/dist/harness/compaction/compaction.js

@@ -1,3 +1,4 @@
+import { contentText, retryAssistantCall, } from "@earendil-works/pi-ai";
 import { convertToLlm, createBranchSummaryMessage, createCompactionSummaryMessage, createCustomMessage, } from "../messages.js";
 import { buildSessionContext } from "../session/session.js";
 import { CompactionError, err, ok } from "../types.js";
@@ -52,6 +53,31 @@ function getMessageFromEntryForCompaction(entry) {
     }
     return getMessageFromEntry(entry);
 }
+export async function completeSimpleWithRetries(models, model, context, options, retry, callbacks) {
+    return retryAssistantCall(() => models.completeSimple(model, context, options), retry, options.signal, callbacks);
+}
+function combineUsage(first, second) {
+    return {
+        input: first.input + second.input,
+        output: first.output + second.output,
+        cacheRead: first.cacheRead + second.cacheRead,
+        cacheWrite: first.cacheWrite + second.cacheWrite,
+        ...(first.cacheWrite1h !== undefined || second.cacheWrite1h !== undefined
+            ? { cacheWrite1h: (first.cacheWrite1h ?? 0) + (second.cacheWrite1h ?? 0) }
+            : {}),
+        ...(first.reasoning !== undefined || second.reasoning !== undefined
+            ? { reasoning: (first.reasoning ?? 0) + (second.reasoning ?? 0) }
+            : {}),
+        totalTokens: first.totalTokens + second.totalTokens,
+        cost: {
+            input: first.cost.input + second.cost.input,
+            output: first.cost.output + second.cost.output,
+            cacheRead: first.cost.cacheRead + second.cost.cacheRead,
+            cacheWrite: first.cost.cacheWrite + second.cost.cacheWrite,
+            total: first.cost.total + second.cost.total,
+        },
+    };
+}
 /** Default compaction settings used by the harness. */
 export const DEFAULT_COMPACTION_SETTINGS = {
     enabled: true,
@@ -355,7 +381,12 @@ Use this EXACT format:
 
 Keep each section concise. Preserve exact file paths, function names, and error messages.`;
 /** Generate or update a conversation summary for compaction. */
-export async function generateSummary(currentMessages, models, model, reserveTokens, signal, customInstructions, previousSummary, thinkingLevel) {
+export async function generateSummary(currentMessages, models, model, reserveTokens, signal, customInstructions, previousSummary, thinkingLevel, retry, callbacks) {
+    const result = await generateSummaryWithUsage(currentMessages, models, model, reserveTokens, signal, customInstructions, previousSummary, thinkingLevel, retry, callbacks);
+    return result.ok ? ok(result.value.text) : err(result.error);
+}
+/** Generate or update a conversation summary and return its provider usage. */
+export async function generateSummaryWithUsage(currentMessages, models, model, reserveTokens, signal, customInstructions, previousSummary, thinkingLevel, retry, callbacks) {
     const maxTokens = Math.min(Math.floor(0.8 * reserveTokens), model.maxTokens > 0 ? model.maxTokens : Number.POSITIVE_INFINITY);
     let basePrompt = previousSummary ? UPDATE_SUMMARIZATION_PROMPT : SUMMARIZATION_PROMPT;
     if (customInstructions) {
@@ -378,18 +409,15 @@ export async function generateSummary(currentMessages, models, model, reserveTok
     const completionOptions = model.reasoning && thinkingLevel && thinkingLevel !== "off"
         ? { maxTokens, signal, reasoning: thinkingLevel }
         : { maxTokens, signal };
-    const response = await models.completeSimple(model, { systemPrompt: SUMMARIZATION_SYSTEM_PROMPT, messages: summarizationMessages }, completionOptions);
+    const response = await completeSimpleWithRetries(models, model, { systemPrompt: SUMMARIZATION_SYSTEM_PROMPT, messages: summarizationMessages }, completionOptions, retry, callbacks);
     if (response.stopReason === "aborted") {
         return err(new CompactionError("aborted", response.errorMessage || "Summarization aborted"));
     }
     if (response.stopReason === "error") {
         return err(new CompactionError("summarization_failed", `Summarization failed: ${response.errorMessage || "Unknown error"}`));
     }
-    const textContent = response.content
-        .filter((c) => c.type === "text")
-        .map((c) => c.text)
-        .join("\n");
-    return ok(textContent);
+    const textContent = contentText(response.content);
+    return ok({ text: textContent, usage: response.usage });
 }
 /** Prepare session entries for compaction, or return undefined when compaction is not applicable. */
 export function prepareCompaction(pathEntries, settings) {
@@ -408,7 +436,9 @@ export function prepareCompaction(pathEntries, settings) {
     if (prevCompactionIndex >= 0) {
         const prevCompaction = pathEntries[prevCompactionIndex];
         previousSummary = prevCompaction.summary;
-        const firstKeptEntryIndex = pathEntries.findIndex((entry) => entry.id === prevCompaction.firstKeptEntryId);
+        const firstKeptEntryIndex = prevCompaction.firstKeptEntryId
+            ? pathEntries.findIndex((entry) => entry.id === prevCompaction.firstKeptEntryId)
+            : -1;
         boundaryStart = firstKeptEntryIndex >= 0 ? firstKeptEntryIndex : prevCompactionIndex + 1;
     }
     const boundaryEnd = pathEntries.length;
@@ -434,6 +464,12 @@ export function prepareCompaction(pathEntries, settings) {
                 turnPrefixMessages.push(msg);
         }
     }
+    const retainedTail = [];
+    for (let i = cutPoint.firstKeptEntryIndex; i < boundaryEnd; i++) {
+        const msg = getMessageFromEntryForCompaction(pathEntries[i]);
+        if (msg)
+            retainedTail.push(msg);
+    }
     const fileOps = extractFileOperations(messagesToSummarize, pathEntries, prevCompactionIndex);
     if (cutPoint.isSplitTurn) {
         for (const msg of turnPrefixMessages) {
@@ -444,6 +480,7 @@ export function prepareCompaction(pathEntries, settings) {
         firstKeptEntryId,
         messagesToSummarize,
         turnPrefixMessages,
+        retainedTail,
         isSplitTurn: cutPoint.isSplitTurn,
         tokensBefore,
         previousSummary,
@@ -467,28 +504,37 @@ Summarize the prefix to provide context for the retained suffix:
 Be concise. Focus on what's needed to understand the kept suffix.`;
 export { serializeConversation } from "./utils.js";
 /** Generate compaction summary data from prepared session history. */
-export async function compact(preparation, models, model, customInstructions, signal, thinkingLevel) {
-    const { firstKeptEntryId, messagesToSummarize, turnPrefixMessages, isSplitTurn, tokensBefore, previousSummary, fileOps, settings, } = preparation;
+export async function compact(preparation, models, model, customInstructions, signal, thinkingLevel, retry, callbacks) {
+    const { firstKeptEntryId, messagesToSummarize, turnPrefixMessages, retainedTail, isSplitTurn, tokensBefore, previousSummary, fileOps, settings, } = preparation;
     if (!firstKeptEntryId) {
         return err(new CompactionError("invalid_session", "First kept entry has no UUID - session may need migration"));
     }
     let summary;
+    let summaryUsage;
     if (isSplitTurn && turnPrefixMessages.length > 0) {
-        const historyResult = messagesToSummarize.length > 0
-            ? await generateSummary(messagesToSummarize, models, model, settings.reserveTokens, signal, customInstructions, previousSummary, thinkingLevel)
-            : ok("No prior history.");
-        if (!historyResult.ok)
-            return err(historyResult.error);
-        const turnPrefixResult = await generateTurnPrefixSummary(turnPrefixMessages, models, model, settings.reserveTokens, signal, thinkingLevel);
+        let historyText = "No prior history.";
+        let historyUsage;
+        if (messagesToSummarize.length > 0) {
+            const historyResult = await generateSummaryWithUsage(messagesToSummarize, models, model, settings.reserveTokens, signal, customInstructions, previousSummary, thinkingLevel, retry, callbacks);
+            if (!historyResult.ok)
+                return err(historyResult.error);
+            historyText = historyResult.value.text;
+            historyUsage = historyResult.value.usage;
+        }
+        const turnPrefixResult = await generateTurnPrefixSummary(turnPrefixMessages, models, model, settings.reserveTokens, signal, thinkingLevel, retry, callbacks);
         if (!turnPrefixResult.ok)
             return err(turnPrefixResult.error);
-        summary = `${historyResult.value}\n\n---\n\n**Turn Context (split turn):**\n\n${turnPrefixResult.value}`;
+        summary = `${historyText}\n\n---\n\n**Turn Context (split turn):**\n\n${turnPrefixResult.value.text}`;
+        summaryUsage = historyUsage
+            ? combineUsage(historyUsage, turnPrefixResult.value.usage)
+            : turnPrefixResult.value.usage;
     }
     else {
-        const summaryResult = await generateSummary(messagesToSummarize, models, model, settings.reserveTokens, signal, customInstructions, previousSummary, thinkingLevel);
+        const summaryResult = await generateSummaryWithUsage(messagesToSummarize, models, model, settings.reserveTokens, signal, customInstructions, previousSummary, thinkingLevel, retry, callbacks);
         if (!summaryResult.ok)
             return err(summaryResult.error);
-        summary = summaryResult.value;
+        summary = summaryResult.value.text;
+        summaryUsage = summaryResult.value.usage;
     }
     const { readFiles, modifiedFiles } = computeFileLists(fileOps);
     summary += formatFileOperations(readFiles, modifiedFiles);
@@ -496,10 +542,12 @@ export async function compact(preparation, models, model, customInstructions, si
         summary,
         firstKeptEntryId,
         tokensBefore,
+        usage: summaryUsage,
+        retainedTail,
         details: { readFiles, modifiedFiles },
     });
 }
-async function generateTurnPrefixSummary(messages, models, model, reserveTokens, signal, thinkingLevel) {
+async function generateTurnPrefixSummary(messages, models, model, reserveTokens, signal, thinkingLevel, retry, callbacks) {
     const maxTokens = Math.min(Math.floor(0.5 * reserveTokens), model.maxTokens > 0 ? model.maxTokens : Number.POSITIVE_INFINITY);
     const llmMessages = convertToLlm(messages);
     const conversationText = serializeConversation(llmMessages);
@@ -511,18 +559,19 @@ async function generateTurnPrefixSummary(messages, models, model, reserveTokens,
             timestamp: Date.now(),
         },
     ];
-    const response = await models.completeSimple(model, { systemPrompt: SUMMARIZATION_SYSTEM_PROMPT, messages: summarizationMessages }, model.reasoning && thinkingLevel && thinkingLevel !== "off"
+    const completionOptions = model.reasoning && thinkingLevel && thinkingLevel !== "off"
         ? { maxTokens, signal, reasoning: thinkingLevel }
-        : { maxTokens, signal });
+        : { maxTokens, signal };
+    const response = await completeSimpleWithRetries(models, model, { systemPrompt: SUMMARIZATION_SYSTEM_PROMPT, messages: summarizationMessages }, completionOptions, retry, callbacks);
     if (response.stopReason === "aborted") {
         return err(new CompactionError("aborted", response.errorMessage || "Turn prefix summarization aborted"));
     }
     if (response.stopReason === "error") {
         return err(new CompactionError("summarization_failed", `Turn prefix summarization failed: ${response.errorMessage || "Unknown error"}`));
     }
-    return ok(response.content
-        .filter((c) => c.type === "text")
-        .map((c) => c.text)
-        .join("\n"));
+    return ok({
+        text: contentText(response.content),
+        usage: response.usage,
+    });
 }
 //# sourceMappingURL=compaction.js.map

package/node_modules/@earendil-works/pi-agent-core/dist/harness/compaction/utils.js

@@ -1,3 +1,4 @@
+import { contentText } from "@earendil-works/pi-ai";
 /** Create an empty file-operation accumulator. */
 export function createFileOps() {
     return {
@@ -78,24 +79,15 @@ export function serializeConversation(messages) {
     const parts = [];
     for (const msg of messages) {
         if (msg.role === "user") {
-            const content = typeof msg.content === "string"
-                ? msg.content
-                : msg.content
-                    .filter((c) => c.type === "text")
-                    .map((c) => c.text)
-                    .join("");
+            const content = contentText(msg.content, "");
             if (content)
                 parts.push(`[User]: ${content}`);
         }
         else if (msg.role === "assistant") {
-            const textParts = [];
             const thinkingParts = [];
             const toolCalls = [];
             for (const block of msg.content) {
-                if (block.type === "text") {
-                    textParts.push(block.text);
-                }
-                else if (block.type === "thinking") {
+                if (block.type === "thinking") {
                     thinkingParts.push(block.thinking);
                 }
                 else if (block.type === "toolCall") {
@@ -109,18 +101,15 @@ export function serializeConversation(messages) {
             if (thinkingParts.length > 0) {
                 parts.push(`[Assistant thinking]: ${thinkingParts.join("\n")}`);
             }
-            if (textParts.length > 0) {
-                parts.push(`[Assistant]: ${textParts.join("\n")}`);
+            if (msg.content.some((block) => block.type === "text")) {
+                parts.push(`[Assistant]: ${contentText(msg.content)}`);
             }
             if (toolCalls.length > 0) {
                 parts.push(`[Assistant tool calls]: ${toolCalls.join("; ")}`);
             }
         }
         else if (msg.role === "toolResult") {
-            const content = msg.content
-                .filter((c) => c.type === "text")
-                .map((c) => c.text)
-                .join("");
+            const content = contentText(msg.content, "");
             if (content) {
                 parts.push(`[Tool result]: ${truncateForSummary(content, TOOL_RESULT_MAX_CHARS)}`);
             }