omnius 1.0.267 → 1.0.268

package/dist/index.js

@@ -557292,7 +557292,7 @@ var init_ollama_pool = __esm({
       elasticSpawnCap(gpus) {
         return this.config.maxSpawnedInstances > 0 ? this.config.maxSpawnedInstances : Math.max(0, gpus.length - 1);
       }
-      async getGpusForPlacement(maxAgeMs = 3e3) {
+      async getGpusForPlacement(maxAgeMs = 3e4) {
         const now2 = Date.now();
         if (this.gpuCache && now2 - this.gpuCache.takenAtMs <= maxAgeMs) {
           return this.gpuCache.gpus;
@@ -557366,7 +557366,12 @@ var init_ollama_pool = __esm({
         }
         const gpuUuid = gpu?.uuid || null;
         const gpuIndex = gpu?.index ?? null;
-        const { proc, ready } = await this.spawner({ port, gpuUuid, gpuIndex, config: this.config });
+        const { proc, ready } = await this.spawner({
+          port,
+          gpuUuid,
+          gpuIndex,
+          config: this.config
+        });
         try {
           await ready;
         } catch (err) {
@@ -562511,6 +562516,8 @@ var init_streaming_executor = __esm({
       insertionOrder = [];
       config;
       executeFn = null;
+      /** Resolver for notification-based wait (delay-1: replaces 1ms spin loop) */
+      _notifyResolve = null;
       constructor(config) {
         this.config = {
           maxConcurrent: config?.maxConcurrent ?? 5,
@@ -562588,6 +562595,18 @@ var init_streaming_executor = __esm({
         }
         return results;
       }
+      /** Signal the notification promise (wake waitAll if sleeping) */
+      _signalChange() {
+        const r2 = this._notifyResolve;
+        this._notifyResolve = null;
+        r2?.();
+      }
+      /** Create a new notification promise for waitAll to await */
+      _waitForChange() {
+        return new Promise((resolve70) => {
+          this._notifyResolve = resolve70;
+        });
+      }
       /**
        * Wait for all queued/executing tools to complete.
        */
@@ -562603,7 +562622,7 @@ var init_streaming_executor = __esm({
             this.processQueue();
           } else {
             this.processQueue();
-            await new Promise((r2) => setTimeout(r2, 1));
+            await this._waitForChange();
           }
         }
       }
@@ -562740,6 +562759,7 @@ var init_streaming_executor = __esm({
             break;
           }
         }
+        this._signalChange();
       }
       startExecution(entry) {
         entry.state = "executing";
@@ -563908,7 +563928,11 @@ function captureToolchainVersions() {
   const out = {};
   for (const probe of TOOLCHAIN_PROBES) {
     try {
-      const v = execSync47(probe.cmd, { encoding: "utf8", stdio: ["ignore", "pipe", "ignore"], timeout: 1500 }).trim();
+      const v = execSync47(probe.cmd, {
+        encoding: "utf8",
+        stdio: ["ignore", "pipe", "ignore"],
+        timeout: 500
+      }).trim();
       out[probe.name] = v.split(/\r?\n/)[0].slice(0, 200);
     } catch {
       out[probe.name] = null;
@@ -563929,7 +563953,10 @@ function freeDiskBytes(path12 = "/tmp") {
     const st = statSync35(path12);
     if (!st.isDirectory())
       return -1;
-    const out = execSync47(`df -P -k ${JSON.stringify(path12)} | tail -1`, { encoding: "utf8", timeout: 1500 });
+    const out = execSync47(`df -P -k ${JSON.stringify(path12)} | tail -1`, {
+      encoding: "utf8",
+      timeout: 500
+    });
     const cols = out.trim().split(/\s+/);
     const availKb = parseInt(cols[3] ?? "0", 10);
     if (!Number.isFinite(availKb))
@@ -581407,21 +581434,30 @@ var init_cascadeBackend = __esm({
         } catch (err) {
           this.consecutiveFailures++;
           if (this.isTransientError(err) && this.consecutiveFailures >= this.maxFailures) {
-            const nextIdx = this.findNextAvailableEndpoint();
-            if (nextIdx !== null && nextIdx !== this.activeIndex) {
+            const candidates = this.endpoints.filter((_ep, i2) => i2 !== this.activeIndex && _ep.modelAvailable !== false);
+            if (candidates.length > 0) {
               const from3 = this.endpoints[this.activeIndex];
-              const to = this.endpoints[nextIdx];
               const reason = `${this.consecutiveFailures} consecutive failures: ${err instanceof Error ? err.message : String(err)}`;
-              this.activeIndex = nextIdx;
-              this.consecutiveFailures = 0;
-              this.onSwitch?.(from3, to, reason);
-              const newBackend = this.getActiveBackend();
+              const promises = candidates.map((ep) => {
+                const bk = this.backends.get(ep.url);
+                return bk.chatCompletion(request).then((result) => ({ result, from: ep }));
+              });
               try {
-                const result = await newBackend.chatCompletion(request);
+                const { result, from: from4 } = await Promise.race(promises);
+                const winnerIdx = this.endpoints.indexOf(from4);
+                if (winnerIdx >= 0) {
+                  this.activeIndex = winnerIdx;
+                  this.onSwitch?.(from4, from4, reason);
+                }
+                this.consecutiveFailures = 0;
                 return result;
-              } catch (cascadeErr) {
-                this.consecutiveFailures++;
-                throw cascadeErr;
+              } catch {
+                this.consecutiveFailures += candidates.length;
+                throw err;
+              } finally {
+                for (const p2 of promises)
+                  p2.catch(() => {
+                  });
               }
             }
           }
@@ -581636,7 +581672,7 @@ function createSteeringIngress(input) {
     timestamp: (/* @__PURE__ */ new Date()).toISOString()
   };
 }
-async function interpretSteeringIngress(backend, ingress, timeoutMs = 15e3) {
+async function interpretSteeringIngress(backend, ingress, timeoutMs = 8e3) {
   const first2 = await backend.chatCompletion({
     messages: [
       {
@@ -594350,6 +594386,175 @@ var init_task_complete_box = __esm({
   }
 });
 
+// packages/cli/src/tui/syntax-highlight.ts
+var syntax_highlight_exports = {};
+__export(syntax_highlight_exports, {
+  detectLanguage: () => detectLanguage2,
+  getHighlightStatus: () => getHighlightStatus,
+  highlightBlock: () => highlightBlock,
+  highlightCode: () => highlightCode,
+  isAvailable: () => isAvailable,
+  prewarm: () => prewarm
+});
+function highlightingDisabled() {
+  return !isTTY2 || noColorEnv || disableEnv;
+}
+async function loadHighlighter() {
+  if (_state.attempted) return _state.fn;
+  _state.attempted = true;
+  if (highlightingDisabled()) {
+    _state.reason = !isTTY2 ? "non-tty" : noColorEnv ? "NO_COLOR set" : "OMNIUS_TUI_HIGHLIGHT=0";
+    return null;
+  }
+  try {
+    const { createRequire: createRequire10 } = await import("node:module");
+    const req3 = createRequire10(import.meta.url);
+    let resolved = null;
+    try {
+      resolved = req3.resolve("cli-highlight");
+    } catch {
+      _state.reason = "cli-highlight not installed";
+      return null;
+    }
+    const mod3 = await import(resolved).catch(() => null);
+    if (!mod3) {
+      _state.reason = "cli-highlight failed to load";
+      return null;
+    }
+    const m2 = mod3;
+    const candidate = m2.highlight ?? m2.default ?? null;
+    if (typeof candidate !== "function") {
+      _state.reason = "cli-highlight export shape unrecognized";
+      return null;
+    }
+    _state.fn = candidate;
+    return candidate;
+  } catch (err) {
+    _state.reason = `import threw: ${err?.message ?? String(err)}`;
+    return null;
+  }
+}
+function loadHighlighterSync() {
+  if (highlightingDisabled()) return null;
+  return _state.fn;
+}
+async function prewarm() {
+  await loadHighlighter();
+}
+function isAvailable() {
+  if (highlightingDisabled()) return false;
+  return _state.attempted && _state.fn !== null;
+}
+function getHighlightStatus() {
+  return {
+    available: isAvailable(),
+    attempted: _state.attempted,
+    reason: _state.reason,
+    isTTY: isTTY2,
+    noColor: noColorEnv,
+    disabledByEnv: disableEnv
+  };
+}
+function highlightCode(code8, language) {
+  if (!code8) return code8;
+  const fn = loadHighlighterSync();
+  if (!fn) return code8;
+  const lang = (language ?? detectLanguage2(code8) ?? "").trim();
+  try {
+    if (lang) {
+      return fn(code8, { language: lang, ignoreIllegals: true });
+    }
+    return fn(code8, { ignoreIllegals: true });
+  } catch {
+    return code8;
+  }
+}
+function detectLanguage2(text2) {
+  if (!text2) return null;
+  const trimmed = text2.trimStart();
+  const shebang = trimmed.match(/^#!\s*\/[^\n]+/);
+  if (shebang) {
+    const sb = shebang[0];
+    if (/python/.test(sb)) return "python";
+    if (/(?:^|[\s/])(?:bash|sh|zsh)\b/.test(sb)) return "bash";
+    if (/node/.test(sb)) return "javascript";
+    if (/ruby/.test(sb)) return "ruby";
+    if (/perl/.test(sb)) return "perl";
+  }
+  if (/^[\s\n]*[{[]/.test(trimmed)) {
+    try {
+      JSON.parse(trimmed);
+      return "json";
+    } catch {
+    }
+  }
+  if (/^[-a-zA-Z_][\w-]*:\s/.test(trimmed) && /\n[-a-zA-Z_][\w-]*:\s/.test(trimmed)) {
+    return "yaml";
+  }
+  if (/^(?:async def |def |class |import |from )/.test(trimmed) && /(?::\s*$|->|self\b)/m.test(trimmed)) {
+    return "python";
+  }
+  if (/^(?:import |export |const |let |var |function |class |interface |type |async )/.test(trimmed)) {
+    if (/(:\s*(?:string|number|boolean|any|unknown|void)\b|\binterface\b|\btype\s+\w+\s*=)/.test(trimmed)) {
+      return "typescript";
+    }
+    return "javascript";
+  }
+  if (/(?:^|\n)\s*(?:fn |use |let mut |impl |struct |enum |trait )/.test(trimmed)) {
+    return "rust";
+  }
+  if (/(?:^package \w+|\nfunc \w+\s*\()/.test(trimmed)) {
+    return "go";
+  }
+  if (/^\s*(?:SELECT|INSERT|UPDATE|DELETE|CREATE|ALTER|DROP)\b/i.test(trimmed)) {
+    return "sql";
+  }
+  if (/^[\s\n]*<(?:!DOCTYPE|html|\?xml|\w+)/i.test(trimmed)) {
+    return "html";
+  }
+  if (/^(?:---|\+\+\+|@@|diff )/.test(trimmed)) {
+    return "diff";
+  }
+  if (/^\s*(?:\$\s|sudo |apt |brew |npm |pnpm |yarn |git |docker |kubectl |curl |wget )/.test(trimmed)) {
+    return "bash";
+  }
+  return null;
+}
+function highlightBlock(code8, language) {
+  if (!code8) return [""];
+  const fn = loadHighlighterSync();
+  if (!fn) return code8.split("\n");
+  const lang = (language ?? detectLanguage2(code8) ?? "").trim();
+  try {
+    const out = lang ? fn(code8, { language: lang, ignoreIllegals: true }) : fn(code8, { ignoreIllegals: true });
+    const lines = out.split("\n");
+    const inputLines = code8.split("\n");
+    if (lines.length === inputLines.length) return lines;
+    if (lines.length < inputLines.length) {
+      while (lines.length < inputLines.length) lines.push("");
+    } else {
+      lines.length = inputLines.length;
+    }
+    return lines;
+  } catch {
+    return code8.split("\n");
+  }
+}
+var isTTY2, noColorEnv, disableEnv, _state;
+var init_syntax_highlight = __esm({
+  "packages/cli/src/tui/syntax-highlight.ts"() {
+    "use strict";
+    isTTY2 = process.stdout?.isTTY ?? false;
+    noColorEnv = process.env["NO_COLOR"] !== void 0 && process.env["NO_COLOR"] !== "";
+    disableEnv = process.env["OMNIUS_TUI_HIGHLIGHT"] === "0";
+    _state = {
+      attempted: false,
+      fn: null,
+      reason: ""
+    };
+  }
+});
+
 // packages/cli/src/tui/model-picker.ts
 import { totalmem as totalmem4 } from "node:os";
 function isImageGenModel(name10, family) {
@@ -596058,10 +596263,29 @@ function diffBodyLines(output, verbose) {
   for (const line of lines.slice(0, maxLines)) {
     const m2 = line.match(DIFF_LINE_RE);
     if (m2) {
-      const colored = m2[1] === "+" ? c3.green(line) : c3.red(line);
+      const marker = m2[1];
+      const codeStart = m2.index + m2[0].length;
+      const prefix = line.slice(0, codeStart);
+      const code8 = line.slice(codeStart);
+      let hlCode = code8;
+      if (isAvailable() && code8.trim()) {
+        const hl = highlightCode(code8);
+        if (hl !== code8) hlCode = hl;
+      }
+      const colorSeq = marker === "+" ? "\x1B[32m" : "\x1B[31m";
+      const resetRe = /\x1B\[0m/g;
+      const adjusted = hlCode.replace(resetRe, `\x1B[0m${colorSeq}`);
+      const colored = `${colorSeq}${prefix}${adjusted}\x1B[0m`;
       body.push({ text: colored, mode: "truncate", kind: "plain" });
     } else {
-      body.push({ text: line, mode: "wrap", kind: "dim" });
+      let formatted = line;
+      if (isAvailable() && line.trim()) {
+        const hl = highlightCode(line);
+        formatted = hl !== line ? hl : c3.dim(line);
+      } else {
+        formatted = c3.dim(line);
+      }
+      body.push({ text: formatted, mode: "wrap", kind: "plain" });
     }
   }
   if (lines.length > maxLines) {
@@ -596081,11 +596305,16 @@ function codePreviewLines(output, maxLines) {
   const shown = lines.slice(start2, start2 + maxLines);
   if (shown.length === 0)
     return [{ text: "(empty file)", mode: "wrap", kind: "dim" }];
-  const body = shown.map((line) => ({
-    text: line,
+  let hlLines = null;
+  if (isAvailable()) {
+    const codeBlock = shown.join("\n");
+    const hl = highlightBlock(codeBlock);
+    if (hl.length === shown.length) hlLines = hl;
+  }
+  const body = shown.map((line, i2) => ({
+    text: hlLines ? hlLines[i2] ?? line : line,
     mode: "truncate",
     kind: "plain"
-    // show content in default terminal color, not dim
   }));
   const remaining = lines.length - start2 - shown.length;
   if (remaining > 0) {
@@ -596580,6 +596809,7 @@ var init_render = __esm({
     init_config();
     init_text_selection();
     init_task_complete_box();
+    init_syntax_highlight();
     init_model_picker();
     init_secret_redactor();
     c3 = {
@@ -610028,10 +610258,10 @@ ${CONTENT_BG_SEQ}`);
 // packages/cli/src/tui/tui-select.ts
 import { AsyncLocalStorage as AsyncLocalStorage2 } from "node:async_hooks";
 function ansi3(code8, text2) {
-  return isTTY2 ? `\x1B[${code8}m${text2}\x1B[0m` : text2;
+  return isTTY3 ? `\x1B[${code8}m${text2}\x1B[0m` : text2;
 }
 function fg2563(code8, text2) {
-  return isTTY2 ? `\x1B[38;5;${code8}m${text2}\x1B[0m` : text2;
+  return isTTY3 ? `\x1B[38;5;${code8}m${text2}\x1B[0m` : text2;
 }
 function stripAnsi3(s2) {
   return s2.replace(/\x1B\[[0-9;]*m/g, "");
@@ -610684,14 +610914,14 @@ ${tuiBgSeq()}`);
     }
   });
 }
-var isTTY2, MENU_ACTIVE_GREEN_256, selectColors, nonInteractiveSelectSurface;
+var isTTY3, MENU_ACTIVE_GREEN_256, selectColors, nonInteractiveSelectSurface;
 var init_tui_select = __esm({
   "packages/cli/src/tui/tui-select.ts"() {
     "use strict";
     init_overlay_lock();
     init_theme();
     init_layout2();
-    isTTY2 = process.stdout.isTTY ?? false;
+    isTTY3 = process.stdout.isTTY ?? false;
     MENU_ACTIVE_GREEN_256 = 154;
     selectColors = {
       blue: (t2) => fg2563(39, t2),
@@ -616167,7 +616397,7 @@ var init_format = __esm({
 import { existsSync as existsSync116 } from "node:fs";
 import { extname as extname17, resolve as resolve55 } from "node:path";
 function ansi4(code8, text2) {
-  return isTTY3 ? `\x1B[${code8}m${text2}\x1B[0m` : text2;
+  return isTTY4 ? `\x1B[${code8}m${text2}\x1B[0m` : text2;
 }
 function stripAnsi4(s2) {
   return s2.replace(/\x1B\[[0-9;]*m/g, "");
@@ -616338,13 +616568,13 @@ function showDropPanel(opts) {
     render2();
   });
 }
-var isTTY3, dc;
+var isTTY4, dc;
 var init_drop_panel = __esm({
   "packages/cli/src/tui/drop-panel.ts"() {
     "use strict";
     init_overlay_lock();
     init_layout2();
-    isTTY3 = process.stdout.isTTY ?? false;
+    isTTY4 = process.stdout.isTTY ?? false;
     dc = {
       bold: (t2) => ansi4("1", t2),
       dim: (t2) => ansi4("38;5;250", t2),
@@ -618643,17 +618873,17 @@ import { tmpdir as tmpdir20 } from "node:os";
 import { join as join131 } from "node:path";
 import { execSync as execSync54 } from "node:child_process";
 function isNeovimActive() {
-  return _state !== null && !_state.cleanedUp;
+  return _state2 !== null && !_state2.cleanedUp;
 }
 function isNeovimFocused() {
-  return _state?.focused ?? false;
+  return _state2?.focused ?? false;
 }
 function refocusNeovim() {
-  if (!_state || _state.cleanedUp || _state.focused) return;
-  toggleFocus(_state);
+  if (!_state2 || _state2.cleanedUp || _state2.focused) return;
+  toggleFocus(_state2);
 }
 async function startNeovimMode(opts) {
-  if (_state && !_state.cleanedUp) {
+  if (_state2 && !_state2.cleanedUp) {
     return "Neovim mode is already active. Use /neovim to exit first.";
   }
   let nvimPath;
@@ -618694,7 +618924,7 @@ async function startNeovimMode(opts) {
   const ptyCols = opts.cols;
   const topOffset = opts.topOffset ?? 0;
   const ptyRows = Math.max(5, opts.contentRows);
-  if (isTTY4) {
+  if (isTTY5) {
     const L = layout();
     const bottomBound = L.contentBottom;
     process.stdout.write(
@@ -618743,7 +618973,7 @@ async function startNeovimMode(opts) {
     installedFilteredListeners: [],
     cleanedUp: false
   };
-  _state = state;
+  _state2 = state;
   const toolbarBtns = [];
   {
     const btns = [
@@ -618763,7 +618993,7 @@ async function startNeovimMode(opts) {
     }
   }
   function renderToolbar() {
-    if (!isTTY4) return;
+    if (!isTTY5) return;
     const L = layout();
     const hdrRow = L.headerContent;
     const fg2 = 252;
@@ -618824,7 +619054,7 @@ async function startNeovimMode(opts) {
     stdin.setRawMode(true);
   }
   stdin.resume();
-  if (isTTY4) {
+  if (isTTY5) {
     process.stdout.write("\x1B[?1002h\x1B[?1006h");
   }
   state.stdinHandler = (data) => {
@@ -618891,12 +619121,12 @@ async function startNeovimMode(opts) {
   return null;
 }
 function stopNeovimMode() {
-  if (!_state || _state.cleanedUp) return Promise.resolve();
+  if (!_state2 || _state2.cleanedUp) return Promise.resolve();
   try {
-    _state.pty?.write("\x1B:qa!\r");
+    _state2.pty?.write("\x1B:qa!\r");
   } catch {
   }
-  const s2 = _state;
+  const s2 = _state2;
   return new Promise((resolve70) => {
     setTimeout(() => {
       if (s2 && !s2.cleanedUp) {
@@ -618907,18 +619137,18 @@ function stopNeovimMode() {
   });
 }
 function writeToNeovimOutput(text2) {
-  if (!_state || _state.cleanedUp || !_state.nvim || !_state.outputChanId) return;
+  if (!_state2 || _state2.cleanedUp || !_state2.nvim || !_state2.outputChanId) return;
   const normalized = text2.replace(/(?<!\r)\n/g, "\r\n");
-  _state.nvim.request("nvim_chan_send", [_state.outputChanId, normalized]).catch(() => {
+  _state2.nvim.request("nvim_chan_send", [_state2.outputChanId, normalized]).catch(() => {
   });
 }
 async function notifyNeovimFileChange(filePath) {
-  if (!_state || _state.cleanedUp || !_state.nvim) return;
+  if (!_state2 || _state2.cleanedUp || !_state2.nvim) return;
   try {
-    await _state.nvim.command("checktime");
+    await _state2.nvim.command("checktime");
     if (filePath) {
       const escaped = filePath.replace(/'/g, "''");
-      await _state.nvim.command(
+      await _state2.nvim.command(
         `if bufexists('${escaped}') | execute 'buffer ' . bufnr('${escaped}') | endif`
       );
     }
@@ -618926,10 +619156,10 @@ async function notifyNeovimFileChange(filePath) {
   }
 }
 function resizeNeovim(cols, contentRows) {
-  if (!_state || _state.cleanedUp || !_state.pty) return;
+  if (!_state2 || _state2.cleanedUp || !_state2.pty) return;
   const rows = Math.max(5, contentRows);
   try {
-    _state.pty.resize(cols, rows);
+    _state2.pty.resize(cols, rows);
   } catch {
   }
 }
@@ -619104,7 +619334,7 @@ function doCleanup(state) {
     } catch {
     }
   }
-  _state = null;
+  _state2 = null;
   process.stdout.write(
     `\x1B[?1000l\x1B[?1002l\x1B[?1003l\x1B[?1006l\x1B[?1015l\x1B[?1004l\x1B[?2004l\x1B[1;${termRows()}r`
     // reset scroll region to full terminal
@@ -619123,16 +619353,16 @@ function doCleanup(state) {
   }
   state.opts.onExit?.();
 }
-var isTTY4, PTY_MODE_ENABLE_RE, STDIN_MOUSE_FOCUS_RE, _state;
+var isTTY5, PTY_MODE_ENABLE_RE, STDIN_MOUSE_FOCUS_RE, _state2;
 var init_neovim_mode = __esm({
   "packages/cli/src/tui/neovim-mode.ts"() {
     "use strict";
     init_setup();
     init_layout2();
-    isTTY4 = process.stdout.isTTY ?? false;
+    isTTY5 = process.stdout.isTTY ?? false;
     PTY_MODE_ENABLE_RE = /\x1B\[\?(?:1004|2004)h/g;
     STDIN_MOUSE_FOCUS_RE = /\x1B\[<[\d;]+[Mm]|\x1B\[M[\s\S]{3}|\x1B\[[IO]|\x1BO[ABCD]/g;
-    _state = null;
+    _state2 = null;
   }
 });
 
@@ -644066,7 +644296,7 @@ function setCarouselWriter(writer) {
   chromeWrite2 = writer;
 }
 function fg(code8, text2) {
-  return isTTY5 ? `\x1B[38;5;${code8}m${text2}\x1B[0m` : text2;
+  return isTTY6 ? `\x1B[38;5;${code8}m${text2}\x1B[0m` : text2;
 }
 function displayWidth(str) {
   let w = 0;
@@ -644104,12 +644334,12 @@ function createRow(phraseIndices, speed, direction, bank2) {
   const phrases = phraseIndices.map((i2) => bank2[i2 % bank2.length]);
   return { phrases, offset: 0, speed, direction, renderedPlain: "" };
 }
-var isTTY5, chromeWrite2, PHRASES, Carousel;
+var isTTY6, chromeWrite2, PHRASES, Carousel;
 var init_carousel = __esm({
   "packages/cli/src/tui/carousel.ts"() {
     "use strict";
     init_layout2();
-    isTTY5 = process.stdout.isTTY ?? false;
+    isTTY6 = process.stdout.isTTY ?? false;
     chromeWrite2 = ((data) => {
       process.stdout.write(data);
     });
@@ -644221,7 +644451,7 @@ var init_carousel = __esm({
        * Sets scroll region to row 5+ for all content/readline.
        */
       start() {
-        if (!isTTY5) return 0;
+        if (!isTTY6) return 0;
         this.started = true;
         setHeaderHeight(this.reservedRows);
         const L = layout();
@@ -644253,7 +644483,7 @@ var init_carousel = __esm({
        * Row 4 is left blank as a separator.
        */
       renderFrame() {
-        if (!isTTY5) return;
+        if (!isTTY6) return;
         const L = layout();
         let buf = "\x1B7";
         buf += "\x1B[?7l";
@@ -644330,7 +644560,7 @@ var init_carousel = __esm({
           process.stdout.removeListener("resize", this.resizeHandler);
           this.resizeHandler = null;
         }
-        if (!isTTY5 || !this.started) return;
+        if (!isTTY6 || !this.started) return;
         const L = layout();
         let buf = "\x1B7";
         for (let i2 = 0; i2 < this.reservedRows; i2++) {
@@ -644566,13 +644796,13 @@ function createAnimatedBanner(id, name10, frameBuilders, frameDurationMs, author
     createdAt: (/* @__PURE__ */ new Date()).toISOString()
   };
 }
-var isTTY6, chromeWrite3, MNEMONIC_ADJECTIVES, MNEMONIC_NOUNS, BannerRenderer;
+var isTTY7, chromeWrite3, MNEMONIC_ADJECTIVES, MNEMONIC_NOUNS, BannerRenderer;
 var init_banner = __esm({
   "packages/cli/src/tui/banner.ts"() {
     "use strict";
     init_theme();
     init_layout2();
-    isTTY6 = process.stdout.isTTY ?? false;
+    isTTY7 = process.stdout.isTTY ?? false;
     chromeWrite3 = ((data) => {
       process.stdout.write(data);
     });
@@ -644734,7 +644964,7 @@ var init_banner = __esm({
        * Returns the number of rows reserved (3 banner + 1 separator = 4).
        */
       start() {
-        if (!isTTY6 || !this.currentDesign) return 0;
+        if (!isTTY7 || !this.currentDesign) return 0;
         this.renderCurrentFrame();
         this._resizeHandler = () => {
           setTermSize(process.stdout.rows ?? 24, process.stdout.columns ?? 80);
@@ -644775,7 +645005,7 @@ var init_banner = __esm({
       }
       /** Render the current frame into the top 3 rows (public for refresh callbacks) */
       renderCurrentFrame() {
-        if (!isTTY6 || !this.currentDesign) return;
+        if (!isTTY7 || !this.currentDesign) return;
         const frame = this.currentDesign.frames[this.currentFrame];
         if (!frame) return;
         this.width = termCols();
@@ -645179,175 +645409,6 @@ var init_carousel_descriptors = __esm({
   }
 });
 
-// packages/cli/src/tui/syntax-highlight.ts
-var syntax_highlight_exports = {};
-__export(syntax_highlight_exports, {
-  detectLanguage: () => detectLanguage2,
-  getHighlightStatus: () => getHighlightStatus,
-  highlightBlock: () => highlightBlock,
-  highlightCode: () => highlightCode,
-  isAvailable: () => isAvailable,
-  prewarm: () => prewarm
-});
-function highlightingDisabled() {
-  return !isTTY7 || noColorEnv || disableEnv;
-}
-async function loadHighlighter() {
-  if (_state2.attempted) return _state2.fn;
-  _state2.attempted = true;
-  if (highlightingDisabled()) {
-    _state2.reason = !isTTY7 ? "non-tty" : noColorEnv ? "NO_COLOR set" : "OMNIUS_TUI_HIGHLIGHT=0";
-    return null;
-  }
-  try {
-    const { createRequire: createRequire10 } = await import("node:module");
-    const req3 = createRequire10(import.meta.url);
-    let resolved = null;
-    try {
-      resolved = req3.resolve("cli-highlight");
-    } catch {
-      _state2.reason = "cli-highlight not installed";
-      return null;
-    }
-    const mod3 = await import(resolved).catch(() => null);
-    if (!mod3) {
-      _state2.reason = "cli-highlight failed to load";
-      return null;
-    }
-    const m2 = mod3;
-    const candidate = m2.highlight ?? m2.default ?? null;
-    if (typeof candidate !== "function") {
-      _state2.reason = "cli-highlight export shape unrecognized";
-      return null;
-    }
-    _state2.fn = candidate;
-    return candidate;
-  } catch (err) {
-    _state2.reason = `import threw: ${err?.message ?? String(err)}`;
-    return null;
-  }
-}
-function loadHighlighterSync() {
-  if (highlightingDisabled()) return null;
-  return _state2.fn;
-}
-async function prewarm() {
-  await loadHighlighter();
-}
-function isAvailable() {
-  if (highlightingDisabled()) return false;
-  return _state2.attempted && _state2.fn !== null;
-}
-function getHighlightStatus() {
-  return {
-    available: isAvailable(),
-    attempted: _state2.attempted,
-    reason: _state2.reason,
-    isTTY: isTTY7,
-    noColor: noColorEnv,
-    disabledByEnv: disableEnv
-  };
-}
-function highlightCode(code8, language) {
-  if (!code8) return code8;
-  const fn = loadHighlighterSync();
-  if (!fn) return code8;
-  const lang = (language ?? detectLanguage2(code8) ?? "").trim();
-  try {
-    if (lang) {
-      return fn(code8, { language: lang, ignoreIllegals: true });
-    }
-    return fn(code8, { ignoreIllegals: true });
-  } catch {
-    return code8;
-  }
-}
-function detectLanguage2(text2) {
-  if (!text2) return null;
-  const trimmed = text2.trimStart();
-  const shebang = trimmed.match(/^#!\s*\/[^\n]+/);
-  if (shebang) {
-    const sb = shebang[0];
-    if (/python/.test(sb)) return "python";
-    if (/(?:^|[\s/])(?:bash|sh|zsh)\b/.test(sb)) return "bash";
-    if (/node/.test(sb)) return "javascript";
-    if (/ruby/.test(sb)) return "ruby";
-    if (/perl/.test(sb)) return "perl";
-  }
-  if (/^[\s\n]*[{[]/.test(trimmed)) {
-    try {
-      JSON.parse(trimmed);
-      return "json";
-    } catch {
-    }
-  }
-  if (/^[-a-zA-Z_][\w-]*:\s/.test(trimmed) && /\n[-a-zA-Z_][\w-]*:\s/.test(trimmed)) {
-    return "yaml";
-  }
-  if (/^(?:async def |def |class |import |from )/.test(trimmed) && /(?::\s*$|->|self\b)/m.test(trimmed)) {
-    return "python";
-  }
-  if (/^(?:import |export |const |let |var |function |class |interface |type |async )/.test(trimmed)) {
-    if (/(:\s*(?:string|number|boolean|any|unknown|void)\b|\binterface\b|\btype\s+\w+\s*=)/.test(trimmed)) {
-      return "typescript";
-    }
-    return "javascript";
-  }
-  if (/(?:^|\n)\s*(?:fn |use |let mut |impl |struct |enum |trait )/.test(trimmed)) {
-    return "rust";
-  }
-  if (/(?:^package \w+|\nfunc \w+\s*\()/.test(trimmed)) {
-    return "go";
-  }
-  if (/^\s*(?:SELECT|INSERT|UPDATE|DELETE|CREATE|ALTER|DROP)\b/i.test(trimmed)) {
-    return "sql";
-  }
-  if (/^[\s\n]*<(?:!DOCTYPE|html|\?xml|\w+)/i.test(trimmed)) {
-    return "html";
-  }
-  if (/^(?:---|\+\+\+|@@|diff )/.test(trimmed)) {
-    return "diff";
-  }
-  if (/^\s*(?:\$\s|sudo |apt |brew |npm |pnpm |yarn |git |docker |kubectl |curl |wget )/.test(trimmed)) {
-    return "bash";
-  }
-  return null;
-}
-function highlightBlock(code8, language) {
-  if (!code8) return [""];
-  const fn = loadHighlighterSync();
-  if (!fn) return code8.split("\n");
-  const lang = (language ?? detectLanguage2(code8) ?? "").trim();
-  try {
-    const out = lang ? fn(code8, { language: lang, ignoreIllegals: true }) : fn(code8, { ignoreIllegals: true });
-    const lines = out.split("\n");
-    const inputLines = code8.split("\n");
-    if (lines.length === inputLines.length) return lines;
-    if (lines.length < inputLines.length) {
-      while (lines.length < inputLines.length) lines.push("");
-    } else {
-      lines.length = inputLines.length;
-    }
-    return lines;
-  } catch {
-    return code8.split("\n");
-  }
-}
-var isTTY7, noColorEnv, disableEnv, _state2;
-var init_syntax_highlight = __esm({
-  "packages/cli/src/tui/syntax-highlight.ts"() {
-    "use strict";
-    isTTY7 = process.stdout?.isTTY ?? false;
-    noColorEnv = process.env["NO_COLOR"] !== void 0 && process.env["NO_COLOR"] !== "";
-    disableEnv = process.env["OMNIUS_TUI_HIGHLIGHT"] === "0";
-    _state2 = {
-      attempted: false,
-      fn: null,
-      reason: ""
-    };
-  }
-});
-
 // packages/cli/src/tui/stream-renderer.ts
 function fg2564(code8, text2) {
   return isTTY8 ? `\x1B[38;5;${code8}m${text2}\x1B[0m` : text2;
@@ -704325,28 +704386,32 @@ ${entry.fullContent}`
           }
           if (lm.intervention) {
             contentWrite(() => {
-              const lines = [`⚠ ${lm.intervention}`];
-              if (lm.details && showAdversary) {
-                const det = String(lm.details);
-                lines.push(...det.split("\n").filter(Boolean));
-              }
-              renderBoxedBlock({
-                title: "Adversary",
-                lines,
-                kind: "plain"
-              });
+              renderWarning(`Adversary: ${lm.intervention}`);
             });
           } else if (lm.details && showAdversary) {
             contentWrite(() => {
-              renderBoxedBlock({
-                title: "Adversary",
-                lines: String(lm.details).split("\n").filter(Boolean),
-                kind: "dim"
-              });
+              renderInfo(
+                `Adversary details: ${String(lm.details).split("\n").filter(Boolean).join(" | ")}`
+              );
             });
           }
         }
         break;
+      case "adversary_reaction":
+        if (event.adversary) {
+          const adv = event.adversary;
+          const conf = adv.confidence != null ? ` (${Math.round(adv.confidence * 100)}%)` : "";
+          const text2 = adv.shortText + conf;
+          contentWrite(() => {
+            renderWarning(`Adversary ${adv.class}: ${text2}`);
+          });
+          if (adv.details) {
+            adversaryBuffer.push(adv.details);
+            if (adversaryBuffer.length > 50)
+              adversaryBuffer.splice(0, adversaryBuffer.length - 50);
+          }
+        }
+        break;
     }
   });
   try {

package/docs/operations/delay-analysis.md

@@ -0,0 +1,230 @@
+# Unnecessary Causes of Delays Between Agent Actions
+
+Analysis of `packages/orchestrator/src/` (95 files) — identified delay sources ranked by impact.
+
+---
+
+## HIGH IMPACT
+
+### 1. `streaming-executor.ts:171-192` — `waitAll()` busy-poll loop
+
+```typescript
+async waitAll(): Promise<void> {
+  while (true) {
+    const pending = Array.from(this.tools.values()).filter(
+      e => e.state === "queued" || e.state === "executing"
+    );
+    if (pending.length === 0) break;
+    const executing = pending.filter(e => e.promise);
+    if (executing.length > 0) {
+      await Promise.allSettled(executing.map(e => e.promise!));
+      this.processQueue();
+    } else {
+      this.processQueue();
+      await new Promise(r => setTimeout(r, 1));  // ← 1ms spin loop
+    }
+  }
+}
+```
+
+**Problem:** When tools are queued but none have started (no `promise`), the loop spins with 1ms `setTimeout` between `processQueue()` calls. Each iteration costs at least 1ms of wall-clock delay. With many queued-but-blocked tools, this adds up.
+
+**Fix:** Replace with Promise-based notification (event emitter or `AbortController`-based wait) so the loop yields until a tool completes or a new tool is enqueued.
+
+---
+
+### 2. `ollama-pool.ts:850-869` — Instance readiness probe loop
+
+```typescript
+await new Promise((r) => setTimeout(r, 500));  // ← 500ms between probes
+```
+
+**Problem:** When spawning a new Ollama instance, the code polls `/api/version` every **500ms** with a 2-second abort timeout. This adds **500-2000ms** of delay per new model instance. If the pool is empty and a model isn't loaded, every first request pays this cost.
+
+**Fix:** Use exponential backoff (100ms → 200ms → 400ms) or a readiness event emitter instead of fixed polling.
+
+---
+
+### 3. `cascadeBackend.ts:105-156` — Sequential fallback chain
+
+```typescript
+const result = await backend.chatCompletion(request);
+// On failure → try newBackend.chatCompletion(request);
+// On failure → try primaryBackend.chatCompletion({ timeoutMs: 10_000 });
+```
+
+**Problem:** The cascade backend tries multiple backends in sequence. Each failure triggers a retry on the next backend. If the first backend is slow to fail (not crash), the full timeout (10s+) is paid before the fallback activates.
+
+**Fix:** Start all backends simultaneously with `Promise.race()` and cancel losers on first success.
+
+---
+
+## MEDIUM IMPACT
+
+### 4. `ollama-pool.ts:1190` — Unbounded slot waiter queue
+
+```typescript
+await new Promise<void>((resolve) => this.slotWaiters.push(resolve));
+```
+
+**Problem:** When all GPU slots are occupied, new requests queue as promises in `slotWaiters`. They only resolve when a slot frees up — but there's no timeout or progress feedback. Under heavy load, this can stall indefinitely.
+
+**Fix:** Add a configurable timeout (e.g., 30s) and reject with a clear "no GPU available" error.
+
+---
+
+### 5. `ollama-pool.ts:1336` — GPU detection on every acquire
+
+```typescript
+const rawGpus = await this.gpuDetector();
+```
+
+**Problem:** GPU detection runs on every placement decision. If `gpuDetector()` calls `nvidia-smi` or similar system commands, this adds **10-100ms** per request.
+
+**Fix:** Cache GPU detection results with a TTL (e.g., 30s). GPU topology rarely changes at runtime.
+
+---
+
+### 6. `tool-batching.ts:247-253` — Forced serial execution of non-concurrent tools
+
+```typescript
+for (const call of batch.calls) {
+  results.push(await executeFn(call));
+}
+```
+
+**Problem:** Write/shell tools are forced to run serially. If the agent produces 3+ write tools in one turn, they queue sequentially. No parallelism for independent writes.
+
+**Fix:** Add a write-conflict analyzer (e.g., file-level locking) to allow parallel writes to different files.
+
+---
+
+### 7. `steeringIntake.ts:97` — 15s steering timeout
+
+```typescript
+timeoutMs = 15_000,
+```
+
+**Problem:** Steering intake has a 15-second timeout. If the model is slow or the prompt is large, this adds a full 15s delay before the steering response arrives.
+
+**Fix:** Reduce to 5-8s for steering (which is advisory, not critical) or make it non-blocking with a best-effort result.
+
+---
+
+### 8. `verifierRunner.ts:161` — 60s test timeout
+
+```typescript
+timeout: 60_000,
+```
+
+**Problem:** Test execution has a 60-second timeout. If tests hang or are slow, this blocks the entire agent loop.
+
+**Fix:** Add a per-test timeout and a total timeout with early termination on repeated failures.
+
+---
+
+## LOW-MEDIUM IMPACT
+
+### 9. `ollama-pool.ts:1236,1257` — VRAM estimation per model
+
+```typescript
+const vramNeededMB = await this.estimateModelVramMB(model);
+```
+
+**Problem:** VRAM estimation runs before every spawn/placement. If it involves model metadata lookups or API calls, this adds latency.
+
+**Fix:** Cache VRAM estimates keyed by model name with a TTL.
+
+---
+
+### 10. `streaming-executor.ts:21-26` — `stableValueKey` deep serialization
+
+```typescript
+function stableValueKey(value: unknown): string {
+  if (value === null || typeof value !== "object") return JSON.stringify(value);
+  if (Array.isArray(value)) return `[${value.map(stableValueKey).join(",")}]`;
+  const record = value as Record<string, unknown>;
+  return `{${Object.keys(record).sort().map((key) => ...).join(",")}}`;
+}
+```
+
+**Problem:** Deep serialization of tool arguments for deduplication. For large args (e.g., file contents), this adds O(n) serialization cost per tool call.
+
+**Fix:** Add a size cap (e.g., 10KB) — if the value exceeds it, use a hash instead of deep serialization.
+
+---
+
+### 11. `streaming-executor.ts:263-291` — Duplicate detection overhead
+
+```typescript
+private findPriorEquivalent(entry: StreamingToolEntry): StreamingToolEntry | null
+private cloneDuplicateResult(entry: StreamingToolEntry): boolean
+private mirrorPriorEquivalent(entry: StreamingToolEntry): boolean
+```
+
+**Problem:** Every tool call is checked against all prior calls for equivalence. This is O(n²) in the number of tools per turn.
+
+**Fix:** Use a hash-based index (tool name + arg hash) for O(1) lookup instead of linear scan.
+
+---
+
+### 12. `ollama-pool.ts:1697` — Stale process cleanup timer
+
+```typescript
+const handle = setTimeout(async () => {
+  const { cleanupStaleOllamaProcesses } = await import("./ollama-pool-cleanup.js");
+  const report = await cleanupStaleOllamaProcesses({ ... });
+}, ...);
+```
+
+**Problem:** Cleanup runs as a deferred `setTimeout` — adds latency before stale processes are actually cleaned up.
+
+**Fix:** Run cleanup eagerly when a slot is freed, not on a timer.
+
+---
+
+### 13. `ollama-pool.ts:298,304` — `execSync` with 3s timeout
+
+```typescript
+{ encoding: "utf8", timeout: 3_000 },
+```
+
+**Problem:** Synchronous child processes block the event loop. If the command takes the full 3s, the entire process stalls.
+
+**Fix:** Use `execFile` with `Promise` wrapper or `spawn` with timeout signal.
+
+---
+
+### 14. `preflightSnapshot.ts:284,311` — Multiple 1.5s probe timeouts
+
+```typescript
+timeout: 1500
+```
+
+**Problem:** Preflight checks run multiple probes, each with 1.5s timeout. If multiple probes fail, this adds 3-6s of startup delay.
+
+**Fix:** Run probes in parallel with `Promise.allSettled()` and use the first success.
+
+---
+
+### 15. `tool-batching.ts:212-240` — Concurrency limit on reads
+
+```typescript
+export async function withConcurrencyLimit<T>(...) {
+  // Uses a limiter to cap parallel reads
+}
+```
+
+**Problem:** Concurrent-safe tools (reads) are limited by a concurrency cap. If the limit is too low, parallel reads are serialized unnecessarily.
+
+**Fix:** Make the concurrency limit configurable and increase the default (e.g., from 2 to 8).
+
+---
+
+## Summary: Top 3 Fixes for Immediate Impact
+
+| Priority | File | Change | Expected Savings |
+|----------|------|--------|------------------|
+| 1 | `streaming-executor.ts` | Replace 1ms spin loop with Promise notification | Eliminates busy-poll latency entirely |
+| 2 | `ollama-pool.ts` | Cache GPU detection with 30s TTL | Eliminates 10-100ms per request |
+| 3 | `cascadeBackend.ts` | Parallelize backend fallbacks with `Promise.race()` | Eliminates cascading timeout accumulation |

package/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "omnius",
-  "version": "1.0.267",
+  "version": "1.0.268",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "omnius",
-      "version": "1.0.267",
+      "version": "1.0.268",
       "bundleDependencies": [
         "image-to-ascii"
       ],

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "omnius",
-  "version": "1.0.267",
+  "version": "1.0.268",
   "description": "AI coding agent powered by open-source models (Ollama/vLLM) — interactive TUI with agentic tool-calling loop",
   "type": "module",
   "main": "./dist/index.js",