@agent-native/core 0.47.1 → 0.48.2

package/dist/agent/production-agent.js

@@ -9,7 +9,7 @@ import { PROVIDER_TO_ENV } from "./engine/provider-env-vars.js";
 import { readAppState } from "../application-state/script-helpers.js";
 import { isDemoModeEnabled } from "../demo/config.js";
 import { redactDemoData, redactDemoString } from "../demo/redact.js";
-import { startRun, subscribeToRun, getActiveRunForThread, getActiveRunForThreadAsync, getRun, abortRun, } from "./run-manager.js";
+import { startRun, subscribeToRun, getActiveRunForThread, getActiveRunForThreadAsync, getRun, abortRun, tryClaimRunSlot, } from "./run-manager.js";
 import { readBody } from "../server/h3-helpers.js";
 import { isReadOnlyShellCommand } from "../coding-tools/index.js";
 import { getRequestRunContext, ensureRequestRunContext, getRequestOrgId, getRequestUserEmail, } from "../server/request-context.js";
@@ -19,7 +19,8 @@ import { createToolSearchEntry, TOOL_SEARCH_ACTION_NAME, } from "./tool-search.j
 import { getDefaultMaxIterations, normalizeMaxIterations, readAgentLoopSettings, } from "./loop-settings.js";
 import { isReasoningEffort, normalizeReasoningEffortForModel, } from "../shared/reasoning-effort.js";
 import { isAgentActionStopError } from "../action.js";
-import { preUploadImageAttachments } from "../file-upload/pre-upload-attachments.js";
+import { writeLedgerEntry, readLedgerEntry, clearLedgerForThread, } from "./run-store.js";
+import { preUploadAttachments } from "../file-upload/pre-upload-attachments.js";
 import { extensionIdFromPathname } from "../extensions/path.js";
 import { applyContextDirectives } from "./context-xray/apply-directives.js";
 import { completeRun as completeProgressRun, startRun as startProgressRun, updateRunProgress, } from "../progress/registry.js";
@@ -450,8 +451,10 @@ function generateRunId() {
 function toolInputActivityLabel(toolName) {
     return toolName ? `Preparing ${toolName} action` : "Preparing action input";
 }
-/** Check if an error is transient and should be retried */
-function isContextTooLongError(err) {
+/** Check if an error is transient and should be retried
+ * @internal exported for unit tests only
+ */
+export function isContextTooLongError(err) {
     if (!(err instanceof Error))
         return false;
     const msg = err.message.toLowerCase();
@@ -459,7 +462,10 @@ function isContextTooLongError(err) {
         msg.includes("input_too_long") ||
         msg.includes("too many tokens") ||
         msg.includes("prompt is too long") ||
-        msg.includes("reduce the length"))
+        msg.includes("reduce the length") ||
+        // Gemini phrasing
+        msg.includes("input token count exceeds") ||
+        msg.includes("request too large"))
         return true;
     if (err instanceof EngineError) {
         const code = (err.errorCode ?? "").toLowerCase();
@@ -468,25 +474,48 @@ function isContextTooLongError(err) {
     }
     return false;
 }
-function isRetryableError(err) {
+/** @internal exported for unit tests only */
+export function isRetryableError(err) {
     if (!(err instanceof Error))
         return false;
     const msg = err.message.toLowerCase();
-    const code = err instanceof EngineError ? (err.errorCode ?? "").toLowerCase() : "";
+    const engineErr = err instanceof EngineError ? err : null;
+    const code = (engineErr?.errorCode ?? "").toLowerCase();
+    // Hard non-retryable codes — check these first.
     if (code === "builder_gateway_timeout")
         return false;
     if (code === "rate_limit_exceeded" ||
         msg.includes("daily gateway request cap"))
         return false;
+    // Prefer structured fields from the engine before falling back to message
+    // keyword matching — avoids false positives on user-supplied text that
+    // happens to contain "rate_limit" etc.
+    if (engineErr) {
+        // Provider explicitly said it is retryable.
+        if (engineErr.providerRetryable === true)
+            return true;
+        // HTTP status-code checks (429, 500, 502, 503, 529 = Anthropic overloaded).
+        const sc = engineErr.statusCode;
+        if (sc === 429 || sc === 500 || sc === 502 || sc === 503 || sc === 529)
+            return true;
+    }
     return (code === "builder_gateway_error" ||
         code === "builder_gateway_network_error" ||
+        code === "http_500" ||
         code === "http_502" ||
         code === "http_503" ||
         code === "http_504" ||
         code === "timeout" ||
+        // Anthropic
         msg.includes("overloaded") ||
         msg.includes("rate_limit") ||
         msg.includes("529") ||
+        // OpenAI phrasing
+        msg.includes("rate limit reached") ||
+        // Google / Gemini
+        msg.includes("resource_exhausted") ||
+        msg.includes("quota exceeded") ||
+        // Generic HTTP codes
         msg.includes("502") ||
         msg.includes("503") ||
         msg.includes("504") ||
@@ -499,6 +528,50 @@ function isRetryableError(err) {
         msg.includes("inactivity timeout") ||
         msg.includes("too much time has passed without sending any data"));
 }
+// ---------------------------------------------------------------------------
+// Context-window overflow recovery
+// ---------------------------------------------------------------------------
+/**
+ * Number of recent messages to protect when trimming tool results.
+ * Messages in the tail of this length are left completely intact; only
+ * older messages have their tool-result text replaced with a stub.
+ */
+const CONTEXT_TRIM_KEEP_TAIL = 10;
+const CONTEXT_TRIM_STUB = "[result trimmed to save context — re-run the tool if needed]";
+/**
+ * Attempt one aggressive trim of old tool-result content to reduce the
+ * context window usage.  Tool-result messages older than the last
+ * {@link CONTEXT_TRIM_KEEP_TAIL} messages have their text replaced with a
+ * short stub.  All user/assistant text messages and the recent tail are
+ * preserved exactly.
+ *
+ * Returns a new array — the original is not mutated.
+ * Returns `null` when there are no trimable tool results (so the caller can
+ * skip the retry).
+ */
+export function trimOldToolResults(messages, keepTail = CONTEXT_TRIM_KEEP_TAIL) {
+    const cutoff = Math.max(0, messages.length - keepTail);
+    let trimmed = false;
+    const result = messages.map((msg, idx) => {
+        // Keep messages in the protected tail intact
+        if (idx >= cutoff)
+            return msg;
+        // Only touch user messages that contain tool-result parts
+        if (msg.role !== "user")
+            return msg;
+        const hasToolResult = msg.content.some((p) => p.type === "tool-result");
+        if (!hasToolResult)
+            return msg;
+        const stubbedContent = msg.content.map((p) => {
+            if (p.type !== "tool-result")
+                return p;
+            trimmed = true;
+            return { ...p, content: CONTEXT_TRIM_STUB };
+        });
+        return { role: "user", content: stubbedContent };
+    });
+    return trimmed ? result : null;
+}
 /** Wait with exponential backoff, respecting abort signal */
 function retryDelay(attempt, signal) {
     const baseMs = RETRY_BASE_DELAY_MS * Math.pow(2, attempt);
@@ -535,16 +608,19 @@ function unwrapTextAttachmentEnvelope(text) {
     const match = text.match(/^<attachment\b[^>]*>\n([\s\S]*)\n<\/attachment>$/);
     return match ? match[1] : text;
 }
-function truncateTextAttachment(text) {
+function truncateTextAttachment(text, attachmentName) {
     if (text.length <= MAX_TEXT_ATTACHMENT_CHARS)
         return text;
     const omitted = text.length - MAX_TEXT_ATTACHMENT_CHARS;
-    return `${text.slice(0, MAX_TEXT_ATTACHMENT_CHARS)}\n\n[Attachment truncated after ${MAX_TEXT_ATTACHMENT_CHARS.toLocaleString()} characters; ${omitted.toLocaleString()} characters omitted to keep the agent request within model context limits.]`;
+    const readHint = attachmentName
+        ? ` Use the \`read-attachment\` tool with name="${escapeAttachmentAttribute(attachmentName)}" to read the rest.`
+        : "";
+    return `${text.slice(0, MAX_TEXT_ATTACHMENT_CHARS)}\n\n[Attachment truncated after ${MAX_TEXT_ATTACHMENT_CHARS.toLocaleString()} characters; ${omitted.toLocaleString()} characters omitted.${readHint}]`;
 }
 function formatTextAttachment(att) {
     if (typeof att.text !== "string" || att.text.length === 0)
         return null;
-    const text = truncateTextAttachment(unwrapTextAttachmentEnvelope(att.text));
+    const text = truncateTextAttachment(unwrapTextAttachmentEnvelope(att.text), att.name);
     const attrs = [
         `name="${escapeAttachmentAttribute(att.name || "attachment")}"`,
         att.contentType
@@ -571,7 +647,20 @@ export function buildUserContentWithAttachments(opts) {
     const userContent = [];
     const textAttachments = [];
     for (const att of opts.attachments ?? []) {
-        if (att.type === "image" && att.data) {
+        if (att.type === "image") {
+            // Prefer the hosted URL when one exists (set by preUploadAttachments).
+            // Anthropic / AI SDK accept URL image parts natively; for other engines
+            // the translate layer falls back to base64 automatically.
+            const uploadedUrl = att.url;
+            if (uploadedUrl) {
+                userContent.push({
+                    type: "image",
+                    url: uploadedUrl,
+                });
+                continue;
+            }
+            if (!att.data)
+                continue;
             const match = att.data.match(/^data:(image\/[^;]+);base64,(.+)$/);
             if (match && isSupportedImageMediaType(match[1])) {
                 userContent.push({
@@ -580,6 +669,17 @@ export function buildUserContentWithAttachments(opts) {
                     mediaType: match[1],
                 });
             }
+            else {
+                // The client sent an image in an unsupported format (HEIC, TIFF, AVIF,
+                // etc.). Inject a short text placeholder so the model knows the image
+                // was present but could not be processed, rather than silently omitting
+                // it and leaving the model confused ("I don't see an image").
+                const mime = match?.[1] ?? att.contentType ?? "unknown format";
+                const label = att.name ? `"${att.name}"` : "An image";
+                textAttachments.push(`[${label} could not be processed — unsupported image format (${mime}). ` +
+                    `Inform the user that only JPEG, PNG, GIF, and WebP images are supported, ` +
+                    `and ask them to convert the file before attaching.]`);
+            }
             continue;
         }
         const filePart = dataUrlToFilePart(att);
@@ -1219,9 +1319,10 @@ export async function runAgentLoop(opts) {
                     }
                     else if (event.type === "thinking-delta") {
                         thinkingBuffer += event.text;
-                        // Thinking deltas are not forwarded to the SSE client yet —
-                        // we accumulate them. In a future iteration, we can surface
-                        // them as a collapsible "reasoning" section in the UI.
+                        // Forward thinking deltas as a distinct event type so the UI
+                        // can render a collapsible "Thinking…" cell while the model
+                        // reasons, then collapse it when content arrives.
+                        send({ type: "thinking", text: event.text });
                     }
                     else if (event.type === "tool-input-start") {
                         if (event.id && event.name) {
@@ -1259,6 +1360,8 @@ export async function runAgentLoop(opts) {
                             throw new EngineError(event.error ?? "Engine stream error", {
                                 errorCode: event.errorCode,
                                 upgradeUrl: event.upgradeUrl,
+                                statusCode: event.statusCode,
+                                providerRetryable: event.providerRetryable,
                             });
                         }
                     }
@@ -1269,7 +1372,22 @@ export async function runAgentLoop(opts) {
                 if (signal.aborted)
                     throw err;
                 if (isContextTooLongError(err)) {
-                    throw new EngineError("Conversation has grown too long. Start a new conversation to continue.", { errorCode: "context_length_exceeded" });
+                    // ── One-shot recovery: trim old tool results and retry once ────────
+                    // Only attempt recovery on the first overflow (retry === 0) to avoid
+                    // infinite trim loops. On subsequent overflows fall through to the
+                    // terminal error.
+                    if (retry === 0) {
+                        const trimmed = trimOldToolResults(contextMessages);
+                        if (trimmed !== null) {
+                            // Replace the sent messages for this iteration with the trimmed
+                            // version, clear any partial output, and retry immediately
+                            // (no delay — context errors are not transient).
+                            contextMessages = trimmed;
+                            send({ type: "clear" });
+                            continue;
+                        }
+                    }
+                    throw new EngineError("Conversation has grown too long. The agent tried to recover automatically but the context is still too large. You can continue in a new chat, or ask the agent to summarize the conversation and continue.", { errorCode: "context_length_exceeded" });
                 }
                 if (retry < maxRetriesForError(err) && isRetryableError(err)) {
                     // Clear partial text from the failed attempt so the retry
@@ -1445,9 +1563,37 @@ export async function runAgentLoop(opts) {
             // this turn (connection drop mid-execution → agent retries → repeat).
             // A write tool that keeps failing likely has a timeout / large-payload
             // problem; retrying indefinitely creates duplicates and confuses users.
+            //
+            // LEDGER RECOVERY: before applying the give-up budget, check whether the
+            // previous invocation's zombie actually completed and wrote its result to
+            // the durable ledger. If so, return the ledger result without re-executing
+            // (prevents the duplicate side effect) and skip counting it toward the
+            // interruption budget.
             if (!actionEntry.readOnly) {
                 const writeCacheKey = toolCallCacheKey(toolCall.name, toolCall.input);
                 const priorInterruptions = writeToolInterruptions.get(writeCacheKey) ?? 0;
+                if (priorInterruptions > 0 && opts.threadId) {
+                    const ledgerResult = await readLedgerEntry(opts.threadId, writeCacheKey);
+                    if (ledgerResult !== null) {
+                        // Zombie completed — recover the real result without re-executing.
+                        const result = `(Recovered from prior interrupted chunk — action already completed.)\n\n` +
+                            ledgerResult;
+                        send({
+                            type: "tool_start",
+                            tool: toolCall.name,
+                            input: toolCall.input,
+                        });
+                        send({ type: "tool_done", tool: toolCall.name, result });
+                        recordToolResult(result, false);
+                        return {
+                            type: "tool-result",
+                            toolCallId: toolCall.id,
+                            toolName: toolCall.name,
+                            toolInput: wireToolInput,
+                            content: result,
+                        };
+                    }
+                }
                 if (priorInterruptions >= MAX_WRITE_TOOL_INTERRUPTIONS) {
                     const result = `The ${toolCall.name} action was interrupted ${priorInterruptions} time(s) in this session — ` +
                         `likely a connection timeout with a large payload. Please start a new chat and try again, ` +
@@ -1521,14 +1667,42 @@ export async function runAgentLoop(opts) {
             let mcpApp;
             try {
                 const timeoutSignal = AbortSignal.timeout(toolTimeoutMs);
+                // Keep a reference to the action promise so we can attach a zombie-
+                // detection continuation AFTER Promise.race abandons it on run abort.
+                // The promise itself is not awaited here — Promise.race owns the await.
+                const actionPromise = Promise.resolve(actionEntry.run(toolCall.input, {
+                    send,
+                    userEmail: getRequestUserEmail(),
+                    orgId: getRequestOrgId() ?? null,
+                    caller: "tool",
+                    attachments: opts.attachments,
+                    signal,
+                }));
+                // When the run is aborted (soft-timeout / user cancel) while this tool
+                // call is in flight, Promise.race below will throw "Run aborted" and the
+                // action's promise becomes a zombie — it keeps running but its result is
+                // never returned to the loop. If the zombie eventually resolves, write
+                // the result to the durable ledger keyed by (threadId, toolKey) so the
+                // next continuation chunk can recover it instead of re-executing the
+                // side effect.
+                if (opts.threadId && !actionEntry.readOnly) {
+                    const ledgerThreadId = opts.threadId;
+                    const ledgerToolKey = toolCallCacheKey(toolCall.name, toolCall.input);
+                    actionPromise
+                        .then((zombieRaw) => {
+                        const zombieMcp = isMcpActionResult(zombieRaw) ? zombieRaw : null;
+                        const zombieText = zombieMcp ? zombieMcp.text : zombieRaw;
+                        const zombieStr = typeof zombieText === "string"
+                            ? zombieText
+                            : JSON.stringify(zombieText, null, 2);
+                        void writeLedgerEntry(ledgerThreadId, ledgerToolKey, zombieStr);
+                    })
+                        .catch(() => {
+                        // Action errored in the zombie — no result to ledger.
+                    });
+                }
                 const raw = await Promise.race([
-                    actionEntry.run(toolCall.input, {
-                        send,
-                        userEmail: getRequestUserEmail(),
-                        orgId: getRequestOrgId() ?? null,
-                        caller: "tool",
-                        attachments: opts.attachments,
-                    }),
+                    actionPromise,
                     new Promise((_, reject) => {
                         timeoutSignal.addEventListener("abort", () => reject(new Error(`Tool call timed out after ${toolTimeoutMs / 1000} seconds`)));
                     }),
@@ -1690,12 +1864,21 @@ export async function runAgentLoop(opts) {
         await flushParallelBatch();
         messages.push({ role: "user", content: toolResultParts });
         if (requestedActionStop) {
-            send({ type: "text", text: requestedActionStop.message });
+            // TypeScript can't track ??= through async closures; cast to known type.
+            const stop = requestedActionStop;
+            send({ type: "text", text: stop.message });
             break;
         }
     }
-    if (!signal.aborted)
+    if (!signal.aborted) {
         send({ type: "done" });
+        // Clean up any zombie-completion ledger entries for this thread now that
+        // the turn completed normally. If the run was aborted the ledger must stay
+        // intact so the next continuation chunk can still recover from it.
+        if (opts.threadId) {
+            void clearLedgerForThread(opts.threadId).catch(() => { });
+        }
+    }
     return usage;
 }
 function backgroundChatProgressRunId(turnId) {
@@ -1838,19 +2021,23 @@ export function createProductionAgentHandler(options) {
                 requestAttachments = preparedRequest.attachments;
             }
         }
-        // Pre-upload chat image attachments through the framework file-upload
-        // provider (Builder.io by default). The model still sees the base64
-        // multimodal content for vision; we only ADD the hosted URL so the agent
-        // can embed the image in slides / docs / outbound messages.
+        // Pre-upload chat attachments (images AND files/PDFs) through the framework
+        // file-upload provider (Builder.io by default). The model still sees the
+        // base64 multimodal content for the current turn; each uploaded attachment
+        // also gets a hosted `url` injected so the agent can embed it in slides,
+        // docs, or outbound messages, and callers can persist a URL reference
+        // instead of the raw base64.
         //
         // When no provider is configured, leave attachments untouched and inject a
-        // hint recommending Builder.io connect — the model can still see the
-        // image, it just can't reference it as a URL until upload is set up.
-        if (hasAttachments && requestAttachments.some((a) => a.type === "image")) {
+        // hint recommending Builder.io connect — the model can still see images via
+        // base64, and files via their data URL / text, but has no hosted URL.
+        if (hasAttachments &&
+            requestAttachments.some((a) => a.type === "image" || a.type === "file" || a.type === "document")) {
             try {
-                const preUpload = await preUploadImageAttachments({
+                const preUpload = await preUploadAttachments({
                     attachments: requestAttachments,
                     ownerEmail,
+                    includeFiles: true,
                 });
                 if (preUpload.injectedText) {
                     requestMessage = requestMessage
@@ -1859,7 +2046,28 @@ export function createProductionAgentHandler(options) {
                 }
             }
             catch (err) {
-                console.warn("[agent-native] preUploadImageAttachments failed:", err instanceof Error ? err.message : String(err));
+                console.warn("[agent-native] preUploadAttachments failed:", err instanceof Error ? err.message : String(err));
+            }
+        }
+        // Persist text-ish attachments as thread-scoped agent_scratch resources so
+        // the model can page through them with the `read-attachment` tool. We do
+        // this here — before buildUserContentWithAttachments — so the resource IDs
+        // are available when we inject the truncation notice into the text content.
+        const textAttachmentResourceMap = new Map();
+        if (hasAttachments && threadId) {
+            try {
+                const { persistTextAttachmentsAsResources } = await import("../server/attachment-actions.js");
+                const stored = await persistTextAttachmentsAsResources({
+                    attachments: requestAttachments,
+                    threadId,
+                    ownerEmail,
+                });
+                for (const [k, v] of stored) {
+                    textAttachmentResourceMap.set(k, v);
+                }
+            }
+            catch (err) {
+                console.warn("[agent-native] persistTextAttachmentsAsResources failed:", err instanceof Error ? err.message : String(err));
             }
         }
         // When a per-request engine override is specified, resolve the API key
@@ -2175,15 +2383,18 @@ export function createProductionAgentHandler(options) {
             ...historyMessages,
             { role: "user", content: userContent },
         ];
-        // If there's already an active run for this thread, reject with 409 so
-        // the client can queue or wait rather than silently aborting the existing run.
+        // Atomically claim the run slot for this thread. The claim checks SQL for
+        // a live (non-stale) running row so two near-simultaneous POSTs on
+        // different serverless isolates both see the correct state — a plain
+        // read-then-act check races on multi-isolate deployments because both
+        // reads see no running row before either insert commits.
         if (threadId) {
-            const existingRun = await getActiveRunForThreadAsync(threadId);
-            if (existingRun?.status === "running") {
+            const slot = await tryClaimRunSlot(threadId);
+            if (!slot.claimed) {
                 setResponseStatus(event, 409);
                 return {
                     error: "Run already in progress for this thread",
-                    activeRunId: existingRun.runId,
+                    activeRunId: slot.activeRunId,
                 };
             }
         }
@@ -2537,7 +2748,10 @@ export function createProductionAgentHandler(options) {
                     : {}),
             };
             send({ type: "activity", label: "Contacting model" });
-            let loopUsage;
+            // loopUsage is always assigned — either via instrumentAgentLoop or
+            // runAgentLoop before use below. The definite-assignment guard is
+            // conservative because the try/catch makes the control flow non-obvious.
+            let loopUsage = undefined;
             let instrumented = false;
             try {
                 const { getObservabilityConfig, instrumentAgentLoop } = await import("../observability/traces.js");