token-goat 2.9.15 → 2.9.16

package/dist/{token-goat-chunk-SFAS46RE.mjs → token-goat-chunk-EJCY6KHO.mjs}

@@ -5,11 +5,11 @@ import {
   isSupportedOcrLang,
   loadConfig,
   resolveOcrLangs
-} from "./token-goat-chunk-S4XRY446.mjs";
+} from "./token-goat-chunk-H3DXI6VD.mjs";
 import {
   ensureDirSync,
   tokenGoatHome
-} from "./token-goat-chunk-PM76YS22.mjs";
+} from "./token-goat-chunk-3H6FQT62.mjs";
 import {
   fenceUntrustedOcrText
 } from "./token-goat-chunk-ERTXEKB6.mjs";

package/dist/{token-goat-chunk-Z6UXPYJA.mjs → token-goat-chunk-F2ARFHLJ.mjs}

@@ -27,7 +27,7 @@ import {
   reconstructLayout,
   resolveDestPage,
   withPdfDocument
-} from "./token-goat-chunk-NEI4NC54.mjs";
+} from "./token-goat-chunk-LKSXAMJB.mjs";
 import "./token-goat-chunk-AH6QILZM.mjs";
 import "./token-goat-chunk-Y4AFKTHK.mjs";
 import "./token-goat-chunk-GMOUBOX4.mjs";

package/dist/{token-goat-chunk-YKG35VHC.mjs → token-goat-chunk-GIMGT34T.mjs}

@@ -5,7 +5,7 @@ import {
 } from "./token-goat-chunk-3BTK54F3.mjs";
 import {
   filtersFilteredToEmptyNotice
-} from "./token-goat-chunk-PM76YS22.mjs";
+} from "./token-goat-chunk-3H6FQT62.mjs";
 import {
   displaySafeText
 } from "./token-goat-chunk-ERTXEKB6.mjs";

package/dist/{token-goat-chunk-SZWYESBS.mjs → token-goat-chunk-GMXBQS6O.mjs}

@@ -4,7 +4,7 @@ import {
   canRunWrappedShell,
   clipToDeliveryCap,
   deliveredOutputBytes
-} from "./token-goat-chunk-YZX7EFG4.mjs";
+} from "./token-goat-chunk-UXHDBVUR.mjs";
 import {
   DEFAULT_RECONCILE_BUDGET_MS,
   HOOK_EVENTS,
@@ -17,11 +17,13 @@ import {
   readStdinJson,
   reconcileProject,
   storeWebOutput
-} from "./token-goat-chunk-OUGNPMDA.mjs";
+} from "./token-goat-chunk-UMXD2M7O.mjs";
 import {
+  BODY_FIRST_TOOL_RESPONSE_KEYS,
   GENERIC_SERVED_OUTPUT_KEY,
   OUTLINE_MAX_REPLACEMENT_RATIO,
   OUTLINE_MIN_HEADINGS,
+  OUTPUT_FIRST_TOOL_RESPONSE_KEYS,
   WEB_FETCH_KEY_SEP,
   accumulateResidentLines,
   appendDirtyPath,
@@ -33,9 +35,16 @@ import {
   clearCurlDownload,
   compactPathFor,
   computeAdaptiveBudget,
+  contextOutput,
+  denyOutput,
+  emitRewrite,
+  emitRewriteIfChanged,
   enqueueDirtyPathSafe,
   ensureWorkerAlive,
+  estimateTokens,
   extractCompactFromMarker,
+  extractToolResponseField,
+  extractToolResultText,
   fenceUntrusted,
   fenceUntrustedSpans,
   fenceWithMatches,
@@ -47,7 +56,9 @@ import {
   getBashOutputId,
   getContextPressure,
   getCurlDownloadPath,
+  getCwd,
   getFileLineRanges,
+  getFilePath,
   getFileServedOutputs,
   getGlobMatchCount,
   getGrepMatchCount,
@@ -56,31 +67,41 @@ import {
   getSessionBashReruns,
   getSessionFiles,
   getSessionWebFetches,
+  getToolInput,
+  getToolName,
+  getTranscriptPath,
   hasSeenImage,
   hasSessionOutput,
   imageQualifiesForShrink,
   incrementSkillHit,
   indexServedBody,
   installedSkillPath,
+  isMcpErrorResponse,
   isStructuralRewriteAccepted,
+  isTextualContentBlock,
   lastTabContextMatches,
   lineMayCarryResidentSignal,
   listSiblingSessionStates,
   loadSessionCache,
   loadSessionState,
+  makeDedupHintHandlers,
   markCompactStale,
   markCompacted,
   markFileTruncated,
   markHintShown,
   matchesAllowPattern,
   matchesDenyPattern,
+  mcpContentBlocks,
   meetsSavingsFloor,
   metadataEndpointRefusal,
+  normalizePayload,
+  passOutput,
   planMarkdownOutline,
   planServedElisions,
   planSourceSkeleton,
   preToolPathDeclined,
   probeImageMeta,
+  readCopilotHooksOwners,
   readTranscriptTail,
   recordBashOutput,
   recordBashRerun,
@@ -98,13 +119,17 @@ import {
   recordSeenImage,
   recordSymbolRead,
   recordWebFetch,
+  registerHook,
   removeOutstandingAgentSpawn,
   repeatedSkillBodyHint,
+  runHook,
   saveSessionState,
   scanAndRecord,
+  serializeOutput,
   servedRunNotice,
   sessionOutputBodyBytes,
   sessionSidecarPath,
+  sessionStateKey,
   setLastTabContext,
   setTranscriptPath,
   shrinkImage,
@@ -117,76 +142,51 @@ import {
   vscodePathAllowed,
   wasCliReadThisSession,
   wasFileReadThisSession,
-  wasHintShown
-} from "./token-goat-chunk-BL5LNGBG.mjs";
+  wasHintShown,
+  withPreservedBlocks
+} from "./token-goat-chunk-XOHTFRDH.mjs";
 import {
   BASH_OUTPUT_SUBDIR,
-  BODY_FIRST_TOOL_RESPONSE_KEYS,
-  OUTPUT_FIRST_TOOL_RESPONSE_KEYS,
   commandHash,
   compressOutput,
-  contextOutput,
   dedupeConsecutive,
-  denyOutput,
   detectFromCommand,
-  emitRewrite,
-  emitRewriteIfChanged,
-  estimateTokens,
-  extractToolResponseField,
-  extractToolResultText,
   filterByName,
   getBashOutput,
-  getCwd,
-  getFilePath,
   getMonitoringRecallHint,
-  getToolInput,
-  getToolName,
-  getTranscriptPath,
   hasBareBackgroundOrNewline,
   hasUnquotedOperator,
   indexRecallEntry,
   isBashEntryStale,
   isBuildCommand,
-  isMcpErrorResponse,
   isRewriteWorthwhile,
   isScopedGitStatusOrDiffStatCommand,
   isTestRunnerCommand,
-  isTextualContentBlock,
-  makeDedupHintHandlers,
-  mcpContentBlocks,
-  normalizePayload,
-  passOutput,
-  readCopilotHooksOwners,
-  registerHook,
   resolveMinNetSavingsBytes,
-  runHook,
-  serializeOutput,
-  sessionStateKey,
   shlexSplit,
   splitOwnTrailingNotices,
   storeBashOutput,
-  summarizeOutputDelta,
-  withPreservedBlocks
-} from "./token-goat-chunk-YQ7WI2CO.mjs";
-import {
-  extractMarkdownHeadings,
-  formatHeadingTreeParts
-} from "./token-goat-chunk-NDRP4KJQ.mjs";
-import {
-  countSymbols
-} from "./token-goat-chunk-2X2EBBC6.mjs";
+  summarizeOutputDelta
+} from "./token-goat-chunk-5USHO7B6.mjs";
 import {
   recordStat,
   redactSecrets,
   savedTokensFromBytes,
   storeBlob,
   stripAnsiEscapes
-} from "./token-goat-chunk-IVUQLQWN.mjs";
+} from "./token-goat-chunk-DMQ32ZSR.mjs";
+import {
+  extractMarkdownHeadings,
+  formatHeadingTreeParts
+} from "./token-goat-chunk-BFNTLGVA.mjs";
+import {
+  countSymbols
+} from "./token-goat-chunk-63QWJ6UM.mjs";
 import {
   detectLanguage,
   getDb,
   languageHasFlag
-} from "./token-goat-chunk-K7F2BFIK.mjs";
+} from "./token-goat-chunk-R3X5TCPB.mjs";
 import {
   detectHarness,
   envBool,
@@ -196,7 +196,7 @@ import {
   isUnderSystemTemp,
   loadConfig,
   shortFingerprint
-} from "./token-goat-chunk-S4XRY446.mjs";
+} from "./token-goat-chunk-H3DXI6VD.mjs";
 import {
   ENV_KEYS,
   IDENTICAL_READ_MIN_BODY_BYTES,
@@ -215,7 +215,7 @@ import {
   isInsideRoot,
   runGit,
   toKB
-} from "./token-goat-chunk-PM76YS22.mjs";
+} from "./token-goat-chunk-3H6FQT62.mjs";
 import {
   UNTRUSTED_TOOL_TAG,
   displaySafePath,
@@ -945,7 +945,7 @@ init_define_import_meta_env();
 var TRACKED_SKILL = "token-goat";
 var MAX_COMMANDS_SHOWN = 8;
 async function currentCommandNames() {
-  const { buildProgram } = await import("./token-goat-chunk-NDPO7GAH.mjs");
+  const { buildProgram } = await import("./token-goat-chunk-QMZMHQRR.mjs");
   return flattenCommandNames(buildCommandManifest(buildProgram()));
 }
 async function recordSkillVersionSnapshot(sessionId, skillName) {
@@ -7835,12 +7835,77 @@ function extractRgSymbolSearch(cmd) {
 }
 function extractCatJsonPipe(cmd) {
   const m = /^cat\s+(?:"([^"]+)"|'([^']+)'|(\S+))\s*\|\s*jq\b/.exec(cmd);
-  if (!m) return null;
-  const filePath = m[1] ?? m[2] ?? m[3];
-  if (!filePath) return null;
-  if (isTempPath(filePath)) return null;
-  if (!/\.(?:json|yaml|yml|toml)$/i.test(filePath)) return null;
-  return { filePath };
+  if (m) {
+    const filePath = m[1] ?? m[2] ?? m[3];
+    if (filePath && !isTempPath(filePath) && /\.(?:json|yaml|yml|toml)$/i.test(filePath)) {
+      return { filePath, isDirectJq: false };
+    }
+  }
+  const jqDirect = /^jq(?:\s+(?:-[a-zA-Z]+|--[a-zA-Z0-9-]+(?:=\S+)?))*\s+(?:"[^"]*"|'[^']*'|(?:\.[a-zA-Z0-9_.*[\]]+))\s+(?:"([^"]+)"|'([^']+)'|(\S+))\s*$/.exec(cmd);
+  if (jqDirect) {
+    const filePath = jqDirect[1] ?? jqDirect[2] ?? jqDirect[3];
+    if (filePath && !isTempPath(filePath) && /\.(?:json|yaml|yml|toml)$/i.test(filePath)) {
+      return { filePath, isDirectJq: true };
+    }
+  }
+  return null;
+}
+function extractPowerShellJsonPipeline(cmd) {
+  let inner = cmd.trim();
+  const w = POWERSHELL_WRAP_RE.exec(inner);
+  if (w) {
+    inner = (w[1] ?? w[2] ?? "").trim();
+  }
+  if (!/\bConvertFrom-Json\b/i.test(inner)) return null;
+  const parseTarget = (segment) => {
+    const readAll = /\[(?:System\.)?IO\.File\]::ReadAllText\(\s*['"]?([^'")\s]+)['"]?\s*\)/i.exec(segment);
+    if (readAll) {
+      const p = readAll[1];
+      return p && /\.(?:json|txt|log|temp)$/i.test(p) ? p : null;
+    }
+    const tokens = segment.trim().split(/\s+/);
+    const cmdlets = /* @__PURE__ */ new Set(["get-content", "gc", "cat", "type"]);
+    const idx = tokens.findIndex((t) => cmdlets.has(t.toLowerCase()));
+    if (idx === -1) return null;
+    let argIdx = idx + 1;
+    while (argIdx < tokens.length) {
+      const tok = tokens[argIdx];
+      if (!tok) break;
+      if (tok.startsWith("-")) {
+        const flag = tok.toLowerCase();
+        if (flag === "-path" && argIdx + 1 < tokens.length) {
+          const next = tokens[argIdx + 1];
+          if (next && !next.startsWith("-")) {
+            const clean2 = next.replace(/^['"]|['"]$/g, "");
+            return /\.(?:json|txt|log|temp)$/i.test(clean2) ? clean2 : null;
+          }
+        }
+        if ((flag === "-encoding" || flag === "-totalcount") && argIdx + 1 < tokens.length) {
+          argIdx += 2;
+          continue;
+        }
+        argIdx++;
+        continue;
+      }
+      const clean = tok.replace(/^['"(]|['")]$/g, "");
+      return /\.(?:json|txt|log|temp)$/i.test(clean) ? clean : null;
+    }
+    return null;
+  };
+  const pipeIdx = inner.search(/\|\s*ConvertFrom-Json\b/i);
+  if (pipeIdx !== -1) {
+    const upstream = inner.slice(0, pipeIdx).trim();
+    return { filePath: parseTarget(upstream) };
+  }
+  const parenMatch = /ConvertFrom-Json\s*\(([^)]+)\)/i.exec(inner);
+  if (parenMatch) {
+    const inside = parenMatch[1] ?? "";
+    return { filePath: parseTarget(inside) };
+  }
+  if (/ConvertFrom-Json\s*\(/i.test(inner)) {
+    return { filePath: null };
+  }
+  return null;
 }
 function extractWslCatFile(cmd) {
   const wslMatch = /^wsl(?:\s+-d\s+\S+)?\s+bash\s+-c\s+"cat(?:\s+(?:-[a-zA-Z]+|--[a-zA-Z-]+))*\s+\/mnt\/([a-z])\/([^"]*)"/.exec(cmd);
@@ -9532,11 +9597,25 @@ function preBashHandlerInner(event) {
   }
   const catJsonPipe = extractCatJsonPipe(cmd);
   if (catJsonPipe !== null) {
-    const { filePath } = catJsonPipe;
+    const { filePath, isDirectJq } = catJsonPipe;
     const hintPath = displaySafePath(cdStripped ? resolveCdHintPath(rawCmd, filePath, hintCwd) : filePath);
     recordStat("session_hint", 0, 0);
+    const label = isDirectJq ? "`jq`" : "`cat | jq`";
+    return contextOutput(
+      label + ' loads the whole file. Use `token-goat json-query "' + hintPath + '" "<key>"` or `token-goat config-get "' + hintPath + '" KEY_NAME` to slice one value.'
+    );
+  }
+  const psJsonPipe = extractPowerShellJsonPipeline(cmd);
+  if (psJsonPipe !== null) {
+    recordStat("session_hint", 0, 0);
+    if (psJsonPipe.filePath) {
+      const hintPath = displaySafePath(cdStripped ? resolveCdHintPath(rawCmd, psJsonPipe.filePath, hintCwd) : psJsonPipe.filePath);
+      return contextOutput(
+        'PowerShell `ConvertFrom-Json` pipeline detected. Use `token-goat json-query "' + hintPath + '" "<selector>"` to extract fields directly without shell conversion scripts.'
+      );
+    }
     return contextOutput(
-      '`cat | jq` loads the whole file. Use `token-goat config-get "' + hintPath + '" KEY_NAME` or `token-goat section "' + hintPath + '::sectionName"` to slice one value.'
+      'PowerShell `ConvertFrom-Json` pipeline detected. Use `token-goat json-query <file> "<selector>"` or `token-goat web-output` to extract fields directly.'
     );
   }
   const catResult = extractCatFile(cmd);
@@ -9907,6 +9986,23 @@ function recordBashFileReadsForSessionCache(cmd, cwd) {
     return;
   }
 }
+async function maybeEmitLargeUncompressedHint(cmd, output, exitCode, cwd, isUnwrapped, event, ansiResult) {
+  const outputBytes = Buffer.byteLength(output, "utf-8");
+  if (outputBytes < 4096 || process.env["TOKEN_GOAT_BASH_COMPRESS"] === "0" || exitCode !== null && exitCode !== 0 || event.raw["_tg_harness"] !== "vscode" && (isCompressibleSingleCommand(cmd) || !isUnwrapped)) {
+    return null;
+  }
+  const key = `bash-uncompressed-4k:${event.sessionId ?? ""}:${shortFingerprint(cmd)}`;
+  if (wasHintShown(key)) return null;
+  markHintShown(key);
+  recordStat("session_hint", 0, 0);
+  const id = await storeBashOutput(cmd, output, exitCode ?? 0, cwd);
+  const kb = Math.round(outputBytes / 1024);
+  const msg = `[tg] Output was ${kb}KB uncompressed. For large tool outputs, run with 'token-goat compress -c "${displaySafeText(cmd)}"' or inspect via 'token-goat bash-output ${id}'.`;
+  if (ansiResult !== null && ansiResult.hookType === "rewriteOutput") {
+    return emitRewrite(msg + "\n" + fenceUntrusted(ansiResult.updatedOutput, UNTRUSTED_TOOL_TAG), "large uncompressed output hint", void 0, "counted-elsewhere");
+  }
+  return contextOutput(msg);
+}
 async function postBashHandler(event) {
   try {
     const rawCmdRaw = extractCommand(event);
@@ -10012,7 +10108,10 @@ async function postBashHandler(event) {
         const genericElision = await maybeElideServedGenericOutput(cmd, output, exitCode, cwd, cacheMinBytes);
         if (genericElision !== null) return genericElision;
       }
-      return maybeStripAnsiOnly(output) ?? passOutput();
+      const ansiStripped = maybeStripAnsiOnly(output);
+      const largeHint2 = await maybeEmitLargeUncompressedHint(cmd, output, exitCode, cwd, isUnwrapped, event, ansiStripped);
+      if (largeHint2 !== null) return largeHint2;
+      return ansiStripped ?? passOutput();
     }
     if (Buffer.byteLength(output, "utf-8") < cacheMinBytes) return passOutput();
     const cacheKey = isCurlGetCommand(cmd) ? extractCurlUrl(cmd) ?? cmd : stripOutputPipeline(cmd);
@@ -10032,6 +10131,8 @@ async function postBashHandler(event) {
     const curlHtmlFold = maybeFoldCurlHtml(cmd, output, id);
     if (curlHtmlFold !== null) return curlHtmlFold;
     const ansiOnly = maybeStripAnsiOnly(output);
+    const largeHint = await maybeEmitLargeUncompressedHint(cmd, output, exitCode, cwd, isUnwrapped, event, ansiOnly);
+    if (largeHint !== null) return largeHint;
     if (ansiOnly !== null) return ansiOnly;
   } catch {
   }
@@ -11171,7 +11272,7 @@ function buildSubagentBriefing() {
     }
     const tail = [];
     tail.push("");
-    tail.push('Before your first read of any file, check for a token-goat command that returns just what you need and run it instead of a full-file read or wide grep: `token-goat symbol <name>`, `token-goat read "file::symbol"`, `token-goat section "file::<heading>"`. Skipping that check is a violation, not an oversight; the only exemptions are a file under ~200 lines you need whole, a never-indexed file, or an image.');
+    tail.push('Before your first read of any file, check for a token-goat command that returns just what you need and run it instead of a full-file read or wide grep: `token-goat symbol <name>`, `token-goat read "file::symbol"`, `token-goat section "file::<heading>"`, `token-goat json-query file \'a.b.c\'`. Skipping that check is a violation, not an oversight; the only exemptions are a file under ~200 lines you need whole, a never-indexed file, or an image.');
     tail.push("");
     tail.push('Report contract: cite evidence as `file::symbol` or a token-goat recall id (`token-goat mcp-output <id> --full`) rather than pasting bodies; paste a fenced block only when the exact bytes are load-bearing; state every unverified claim explicitly (e.g. "not verified: ...") rather than omitting it.');
     const withCacheIds = head.join("\n") + cacheIdsBlock + "\n" + tail.join("\n");

package/dist/{token-goat-chunk-S4XRY446.mjs → token-goat-chunk-H3DXI6VD.mjs}

@@ -13,7 +13,7 @@ import {
   foldPath,
   projectConfigPath,
   runGit
-} from "./token-goat-chunk-PM76YS22.mjs";
+} from "./token-goat-chunk-3H6FQT62.mjs";
 import {
   normalizeDarwinSystemAlias
 } from "./token-goat-chunk-ERTXEKB6.mjs";
@@ -1421,7 +1421,8 @@ var CONFIG_DEFAULTS = {
   image_shrink: {
     enabled: true,
     jpeg_quality: 75,
-    max_image_pixels: 16e6,
+    // 16,000,000 used to reject a routine 24MP DSLR/phone photo outright -- the exact input this feature exists to shrink -- while the engine's own decode ceiling (MAX_DECODED_BYTES in image_engine.ts, 256MB at 4 bytes/pixel for a single frame) sits at 67,108,864px. 64,000,000 covers cameras well past 24MP (including 48MP phone sensors) with ~4.9% margin below that hard ceiling, which assertDecodableSize enforces independently of this value on every decode regardless of what this header check allows through. Measured on this engine (2026-09-17): a 24MP PNG decodes+shrinks in ~370ms, 60,000,516px in ~920ms, and 67,108,864px (the ceiling itself) in ~760ms -- all comfortably fast, so the new value is chosen for coverage headroom under the ceiling, not for decode-time cost. Decode time is not the cost that binds, though: memory is. This value is read off a header before anything is decoded, so what it really sets is how large a buffer a file's own declared dimensions may talk this hook into allocating -- at 4 bytes per pixel, raising it from 16,000,000 raised that peak from ~64MB to ~256MB, which is why it now sits flush against MAX_DECODED_BYTES rather than comfortably below it. And on the JPEG path there is no assertDecodableSize call at all: decodeJpeg hands the buffer to jpeg-js, so a JPEG decode is bounded by whatever maxResolutionInMP/maxMemoryUsageInMB that call passes -- a bound this codebase states only because decodeJpeg passes both explicitly rather than inheriting jpeg-js's own 100MP default, which is looser than the 67.1MP MAX_DECODED_BYTES allows.
+    max_image_pixels: 64e6,
     screenshot_redirect: true,
     ocr_enabled: true,
     // Confidence is Tesseract's own 0-100 mean-word-confidence score. 65 is a deliberately
@@ -1566,7 +1567,7 @@ var CONFIG_DEFAULTS = {
     offline: false
   },
   mcp: {
-    confine_reads_to_project_root: true,
+    confine_reads_to_project_root: false,
     allowed_roots: []
   },
   hint_stats: {

package/dist/{token-goat-chunk-6DLVZDB6.mjs → token-goat-chunk-LJL6FGDS.mjs}

@@ -4,26 +4,27 @@ import {
   buildEvent,
   relay,
   relayInProcess
-} from "./token-goat-chunk-SZWYESBS.mjs";
-import "./token-goat-chunk-YZX7EFG4.mjs";
+} from "./token-goat-chunk-GMXBQS6O.mjs";
+import "./token-goat-chunk-UXHDBVUR.mjs";
 import {
   MAX_STDIN_BYTES,
   readStdinJson
-} from "./token-goat-chunk-OUGNPMDA.mjs";
-import "./token-goat-chunk-BL5LNGBG.mjs";
-import "./token-goat-chunk-YQ7WI2CO.mjs";
+} from "./token-goat-chunk-UMXD2M7O.mjs";
+import "./token-goat-chunk-XOHTFRDH.mjs";
+import "./token-goat-chunk-5USHO7B6.mjs";
+import "./token-goat-chunk-DMQ32ZSR.mjs";
 import "./token-goat-chunk-3BTK54F3.mjs";
-import "./token-goat-chunk-NDRP4KJQ.mjs";
-import "./token-goat-chunk-2X2EBBC6.mjs";
-import "./token-goat-chunk-Y4AFKTHK.mjs";
-import "./token-goat-chunk-IVUQLQWN.mjs";
-import "./token-goat-chunk-K7F2BFIK.mjs";
-import "./token-goat-chunk-S4XRY446.mjs";
+import "./token-goat-chunk-BFNTLGVA.mjs";
+import "./token-goat-chunk-63QWJ6UM.mjs";
+import "./token-goat-chunk-R3X5TCPB.mjs";
+import "./token-goat-chunk-H3DXI6VD.mjs";
 import "./token-goat-chunk-EEIDFMEM.mjs";
 import "./token-goat-chunk-OSUFN2FV.mjs";
-import "./token-goat-chunk-PM76YS22.mjs";
+import "./token-goat-chunk-3H6FQT62.mjs";
 import "./token-goat-chunk-ERTXEKB6.mjs";
+import "./token-goat-chunk-Y4AFKTHK.mjs";
 import "./token-goat-chunk-GMOUBOX4.mjs";
+import "./token-goat-chunk-RRNZMM3A.mjs";
 import "./token-goat-chunk-A37V4PBF.mjs";
 export {
   MAX_STDIN_BYTES,

package/dist/{token-goat-chunk-NEI4NC54.mjs → token-goat-chunk-LKSXAMJB.mjs}

@@ -280,7 +280,7 @@ async function extractPdfText(data, pagesSpec, layout = false) {
     for (let i = start; i <= end; i++) {
       const page = await getPageWithinDeadline(doc, i, deadline);
       const textItems = await readPageTextItems(page, MAX_PDF_TEXT_BYTES - spent, deadline);
-      const pageText = layout ? reconstructLayout(textItems, deadline) : textItems.map((item) => item.str).join(" ");
+      const pageText = layout ? reconstructLayout(textItems, deadline) : joinTextItems(textItems, "\n");
       spent += pageText.length;
       pages.push(pageText.trim());
     }
@@ -306,7 +306,7 @@ async function locatePdfPages(data, pattern, opts) {
     for (; i <= end && matches.length < maxMatches; i++) {
       const page = await getPageWithinDeadline(doc, i, deadline);
       const textItems = await readPageTextItems(page, MAX_PDF_TEXT_BYTES, deadline);
-      const pageText = textItems.map((item) => item.str).join(" ");
+      const pageText = joinTextItems(textItems, " ");
       const m = re.exec(pageText);
       if (m === null) continue;
       matches.push({ page: i, snippet: locateSnippet(pageText, m.index, m[0].length, context) });
@@ -315,6 +315,11 @@ async function locatePdfPages(data, pattern, opts) {
     return { matches, truncated };
   });
 }
+function joinTextItems(items, eol) {
+  let out = "";
+  for (const item of items) out += item.str + (item.hasEOL ? eol : "");
+  return out;
+}
 function locateSnippet(text, index, matchLen, context) {
   const kept = Math.min(matchLen, context);
   const pad = Math.max(0, context - kept);

package/dist/{token-goat-chunk-A4JYKD5H.mjs → token-goat-chunk-M2R4UQ3B.mjs}

@@ -2,11 +2,11 @@ import { createRequire as __cjsRequire } from 'node:module';
 const require = __cjsRequire(import.meta.url);
 import {
   loadConfig
-} from "./token-goat-chunk-S4XRY446.mjs";
+} from "./token-goat-chunk-H3DXI6VD.mjs";
 import "./token-goat-chunk-EEIDFMEM.mjs";
 import {
   VERSION
-} from "./token-goat-chunk-PM76YS22.mjs";
+} from "./token-goat-chunk-3H6FQT62.mjs";
 import {
   displaySafeJson
 } from "./token-goat-chunk-ERTXEKB6.mjs";

package/dist/{token-goat-chunk-MZDIJJ3R.mjs → token-goat-chunk-MA5BZ2KI.mjs}

@@ -10,13 +10,13 @@ import {
 import {
   createLazyModuleLoader
 } from "./token-goat-chunk-AH6QILZM.mjs";
+import {
+  pushAll
+} from "./token-goat-chunk-3H6FQT62.mjs";
 import {
   DocumentRefusedError,
   MAX_DOCUMENT_WORK_MILLIS
 } from "./token-goat-chunk-Y4AFKTHK.mjs";
-import {
-  pushAll
-} from "./token-goat-chunk-PM76YS22.mjs";
 import {
   init_define_import_meta_env
 } from "./token-goat-chunk-A37V4PBF.mjs";
@@ -332,6 +332,26 @@ function pushTextValue(runs, val) {
     runs.push(String(val["#text"]));
   }
 }
+var MATH_PARA_RE = /<m:oMathPara(?:\s[^>]*)?>([\s\S]*?)<\/m:oMathPara>/g;
+var MATH_RE = /<m:oMath(?:\s[^>]*)?>([\s\S]*?)<\/m:oMath>/g;
+var MATH_TEXT_RE = /<m:t(?:\s[^>]*)?>([^<]*)<\/m:t>/g;
+function mathRunText(fragment) {
+  let out = "";
+  for (const m of fragment.matchAll(MATH_TEXT_RE)) out += m[1];
+  return out;
+}
+function inlineMathRuns(xml, openRun, closeRun) {
+  const asRun = (text) => text.length > 0 ? `${openRun}${text}${closeRun}` : "";
+  const withParas = xml.replace(MATH_PARA_RE, (_whole, inner) => {
+    const lines = [];
+    for (const m of inner.matchAll(MATH_RE)) {
+      const text = mathRunText(m[1]);
+      if (text.length > 0) lines.push(text);
+    }
+    return asRun(lines.length > 0 ? lines.join(" ") : mathRunText(inner));
+  });
+  return withParas.replace(MATH_RE, (_whole, inner) => asRun(mathRunText(inner)));
+}
 function collectTextRuns(node, tag, skipInside = []) {
   const runs = [];
   function walk(n) {
@@ -354,6 +374,21 @@ function collectTextRuns(node, tag, skipInside = []) {
   walk(node);
   return runs;
 }
+function dropAlternateContentFallbacks(node) {
+  if (Array.isArray(node)) {
+    node.forEach(dropAlternateContentFallbacks);
+    return;
+  }
+  if (node === null || typeof node !== "object") return;
+  for (const [key, val] of Object.entries(node)) {
+    if (key === "mc:AlternateContent") {
+      for (const alt of Array.isArray(val) ? val : [val]) {
+        if (alt !== null && typeof alt === "object" && "mc:Choice" in alt && "mc:Fallback" in alt) delete alt["mc:Fallback"];
+      }
+    }
+    if (val !== null && typeof val === "object") dropAlternateContentFallbacks(val);
+  }
+}
 function collectElements(node, tag, opts = {}) {
   const skip = opts.skipInside;
   const out = [];
@@ -413,7 +448,9 @@ export {
   ooxmlPartBudget,
   decodeZipEntry,
   parseOoxmlPart,
+  inlineMathRuns,
   collectTextRuns,
+  dropAlternateContentFallbacks,
   collectElements,
   collectParagraphTexts,
   sortNumberedParts