@tiens.nguyen/gu-cli 1.0.814 → 1.0.817

package/README.md

@@ -10,7 +10,8 @@ gu            # open the REPL in the current folder
 
 The package installs three commands: `gu` (the terminal REPL), `gu-cli` and `gu-local-worker`
 (the same worker CLI). Requires Node.js 18 or newer. The install runs `fix-path` for you; if
-npm's global bin folder is not on your PATH afterwards, run `gu-cli fix-path`.
+npm's global bin folder is not on your PATH afterwards, run `npx @tiens.nguyen/gu-cli fix-path`
+(works even when `gu-cli` itself is not found, e.g. when npm skipped the install script).
 On Windows, PowerShell's built-in `gu` alias (Get-Unique) shadows this CLI — run
 `gu-cli fix-powershell` to drop it.
 

package/client-expert-run.mjs

@@ -40,7 +40,15 @@ import {
   LOCAL_MONGO_URI,
 } from "./client-expert-api.mjs";
 import { askOllamaSetup, localOllamaDefault } from "./ollama-setup.mjs";
-import { SKIP_ALL, createHinter, modelSetupMenu, parseModelSetupChoice } from "./model-setup-ui.mjs";
+import {
+  SKIP_ALL,
+  createHinter,
+  modelSetupMenu,
+  parseModelSetupChoice,
+  providerChoiceMenu,
+  parseProviderChoice,
+} from "./model-setup-ui.mjs";
+import { providerPresets } from "./coding-provider-setup.mjs";
 import { presetFilesIn } from "./presets-transfer.mjs";
 import { readdirSync, statSync, readFileSync } from "node:fs";
 import { parseOllamaTags, formatSize } from "./agent-model-command.mjs";
@@ -312,8 +320,8 @@ export async function ensureNativeMongoMac(ask, mongo, { log = (l) => console.lo
   log(dim(installing ? "\n  It is not installed here. This installs it (Homebrew):\n" : "\n  It is installed but not running. This starts it:\n"));
   for (const s of steps) log(`    brew ${s.join(" ")}`);
   log("");
-  const yes = await ask(installing ? "  Install MongoDB now? [Y/n]: " : "  Start MongoDB now? [Y/n]: ");
-  if (String(yes ?? "").trim().toLowerCase().startsWith("n")) {
+  const yes = await readYesNo(ask, installing ? "  Install MongoDB now? [Y/n]: " : "  Start MongoDB now? [Y/n]: ", true, log);
+  if (!yes) {
     log(dim(`\n  Not ${installing ? "installed" : "started"} — run the commands above yourself.`));
     return { ok: false, explained: true };
   }
@@ -440,8 +448,8 @@ export async function ensureNativeMongo(ask, mongo, {
   const needs = platform === "win32"
     ? "Windows may ask for administrator permission"
     : isRoot ? "" : "sudo will ask for your password";
-  const yes = await ask(`  ${installing ? "Install" : "Start"} MongoDB now?${needs ? ` (${needs})` : ""} [Y/n]: `);
-  if (String(yes ?? "").trim().toLowerCase().startsWith("n")) {
+  const yes = await readYesNo(ask, `  ${installing ? "Install" : "Start"} MongoDB now?${needs ? ` (${needs})` : ""} [Y/n]: `, true, log);
+  if (!yes) {
     log(dim(`\n  Not ${installing ? "installed" : "started"} — run the commands above yourself.`));
     return { ok: false, explained: true };
   }
@@ -608,8 +616,7 @@ export async function startMongoInDocker(ask, { platform = process.platform } =
   console.log(dim(`  Simplest path: ${plan.summary}.`));
   console.log(dim("  Your conversation history lives in that volume and survives a restart."));
   console.log(dim("\n    " + plan.commands[0] + "\n"));
-  const yes = await ask("  Start MongoDB in Docker now? [Y/n]: ");
-  if (String(yes ?? "").trim().toLowerCase().startsWith("n")) return false;
+  if (!(await readYesNo(ask, "  Start MongoDB in Docker now? [Y/n]: ", true))) return false;
 
   for (const cmd of plan.commands) {
     console.log(dim(`  $ ${cmd}`));
@@ -672,13 +679,36 @@ export async function checkMongoConnection(uri) {
  * Returns { uri, custom } to continue with, or { exit: true } when the user declines — the
  * caller ends setup with the `gu clear` → `gu` hint.
  */
+/**
+ * A [Y/n] / [y/N] answer (GitHub #128). Enter → the default; y/yes → yes; n/no → no. Anything else
+ * ("3", "maybe") is NOT taken as an answer: the question is asked again with what to type —
+ * "3" at "Install MongoDB now? [Y/n]" used to install MongoDB. A closed input (null) is "no", as
+ * it always was. After three unrecognised answers the default is used.
+ */
+export function parseYesNo(raw, defaultYes) {
+  if (raw === null || raw === undefined) return false;
+  const a = String(raw).trim().toLowerCase();
+  if (!a) return defaultYes;
+  if (a === "y" || a === "yes") return true;
+  if (a === "n" || a === "no") return false;
+  return null;
+}
+
+export async function readYesNo(ask, prompt, defaultYes, say = (l) => console.log(l)) {
+  for (let attempt = 0; attempt < 3; attempt += 1) {
+    const verdict = parseYesNo(await ask(prompt), defaultYes);
+    if (verdict !== null) return verdict;
+    say(dim(`  Type y or n (Enter = ${defaultYes ? "yes" : "no"}).`));
+  }
+  return defaultYes;
+}
+
 export async function askMongoCredentials(ask, askSecret = null) {
   const { host, port } = mongoHostPort(LOCAL_MONGO_URI);
   console.log(dim(`  MongoDB is answering at ${host}:${port}.`));
   console.log(dim("  If that is a MongoDB this machine already had, it may have its OWN"));
   console.log(dim("  username and password — which setup does not know, and cannot change."));
-  const yes = await ask("  Use the default (empty username and password)? [Y/n]: ");
-  if (!String(yes ?? "").trim().toLowerCase().startsWith("n")) {
+  if (await readYesNo(ask, "  Use the default (empty username and password)? [Y/n]: ", true)) {
     return { uri: LOCAL_MONGO_URI, custom: false };
   }
   // NO → the instance has its own credentials. Ask for them, and CHECK before any installing
@@ -700,15 +730,13 @@ export async function askMongoCredentials(ask, askSecret = null) {
     }
     if (verdict.unverifiable) {
       console.log("\n  " + yellow("⚠ could not verify: ") + dim(verdict.why));
-      const carryOn = await ask("  Continue with these credentials anyway? [Y/n]: ");
-      if (!String(carryOn ?? "").trim().toLowerCase().startsWith("n")) {
+      if (await readYesNo(ask, "  Continue with these credentials anyway? [Y/n]: ", true)) {
         return { uri, custom: true, unverified: true };
       }
     } else {
       console.log("\n  " + red("✗ Check Connection failed") + dim(` — ${verdict.why}`));
     }
-    const again = await ask("  Re-enter the username and password? [y/N]: ");
-    if (!String(again ?? "").trim().toLowerCase().startsWith("y")) return { exit: true };
+    if (!(await readYesNo(ask, "  Re-enter the username and password? [y/N]: ", false))) return { exit: true };
   }
   return { exit: true };
 }
@@ -1304,6 +1332,33 @@ export async function runClientExpertSetup({
       chosenModels = { skipped: true, keys: [], loadPreset: path === "preset" };
       throw new SkipModelQuestion();
     }
+    // WHICH PROVIDER (#132 follow-up): Ollama is still the Enter default, but a hosted provider can be
+    // picked from the preset list or typed. That pick is HANDED to gu's own add-provider flow, which
+    // asks the key and then the model once gu starts — the terminal that flow needs does not exist yet.
+    const provPresets = providerPresets("coding");
+    console.log("");
+    const [ptitle, ...prows] = providerChoiceMenu(provPresets);
+    console.log("  " + bold(ptitle));
+    for (const r of prows) console.log(dim("  " + r));
+    let provider = null;
+    for (let t = 0; t < 3 && !provider; t++) {
+      provider = parseProviderChoice(await askStep("  choose [1]: "), provPresets);
+    }
+    provider = provider ?? { kind: "ollama" };
+    if (provider.kind === "skip") {
+      chosenModels = { skipped: true, keys: [], loadPreset: false };
+      throw new SkipModelQuestion();
+    }
+    if (provider.kind === "preset" || provider.kind === "url") {
+      chosenModels = {
+        skipped: true,
+        keys: [],
+        addProvider: provider.kind === "preset"
+          ? { name: provider.preset.name, url: provider.preset.url, dialect: provider.preset.dialect ?? "openai" }
+          : { name: "", url: "", dialect: "openai" },
+      };
+      throw new SkipModelQuestion();
+    }
     console.log("");
     console.log("  " + bold("Which Ollama server should this machine use?"));
     // THIS MACHINE FIRST, when it has one. Client Expert is the mode whose whole promise is
@@ -1351,6 +1406,7 @@ export async function runClientExpertSetup({
   // #301: skipped on purpose (up front, "s" at both servers, or Esc) → no model field is written.
   const skipModels = Boolean(chosenModels?.skipped);
   const loadPreset = Boolean(chosenModels?.loadPreset);
+  const addProvider = chosenModels?.addProvider ?? null;
 
   try {
     const written = await seedSettings({
@@ -1371,7 +1427,9 @@ export async function runClientExpertSetup({
     }
     if (skipModels) {
       console.log("");
-      console.log("  " + cyan("✓ Model setup skipped") + dim(loadPreset
+      console.log("  " + cyan("✓ Model setup skipped") + dim(addProvider
+        ? ` — ${addProvider.name || "the provider list"}: its key and model are asked as soon as gu starts`
+        : loadPreset
         ? " — the preset picker opens as soon as gu starts"
         : " — set them any time with /model (or /use-presets)"));
     } else if (Object.keys(written).length === 0) {
@@ -1423,5 +1481,5 @@ export async function runClientExpertSetup({
   );
   // #301: the caller opens /use-presets once the terminal exists — the picker, the plan and the
   // confirmation are that command's, not rebuilt here.
-  return { complete: true, stoppedAt: null, ...(loadPreset ? { loadPreset: true } : {}) };
+  return { complete: true, stoppedAt: null, ...(loadPreset ? { loadPreset: true } : {}), ...(addProvider ? { addProvider } : {}) };
 }

package/coding-provider-setup.mjs

@@ -260,6 +260,12 @@ export function providerPresets(role = "coding") {
     { name: "Claude API (Anthropic)", url: CLAUDE_API_URL, roles: ["coding", "agent"] },
     { name: "OpenAI", url: DEFAULT_PROVIDER_URL, roles: ["coding", "agent", "embed"] },
     { name: "Kimi (Moonshot)", url: "https://api.moonshot.ai/v1", roles: ["coding", "agent"] },
+    // #132. Ollama's cloud speaks Ollama's own API (/api/tags), so the preset names its dialect.
+    { name: "Ollama cloud", url: "https://ollama.com", dialect: "ollama", roles: ["coding", "agent"] },
+    // Qwen's two plans take DIFFERENT keys (Token Plan sk-sp-..., pay-as-you-go sk-ws-...), so the
+    // plan is in the name. Both are OpenAI-compatible and already carry their /compatible-mode/v1 path.
+    { name: "Qwen Token Plan (OpenAI-compatible)", url: "https://token-plan.maas.qwencloudapi.com/compatible-mode/v1", roles: ["coding", "agent"] },
+    { name: "Qwen pay-as-you-go (OpenAI-compatible)", url: "https://maas.qwencloudapi.com/compatible-mode/v1", roles: ["coding", "agent"] },
   ];
   return all.filter((p) => p.roles.includes(role));
 }

package/gonext_agent_chat.py

@@ -4574,6 +4574,8 @@ def _strip_tool_tags(text: str) -> str:
     """Remove the internal hint tags we append to tool output (e.g. '[SUCCESS …]',
     '[NOTE: …]', 'Note: This URL failed …') so they never leak into the user reply."""
     out = []
+    # DeepSeek tool_calls/invoke markup must never reach the user as the answer (#133).
+    text = _strip_toolcall_markup(text)
     for ln in (text or "").splitlines():
         s = ln.strip()
         if s.startswith("[SUCCESS") or s.startswith("[NOTE:") or s.startswith("Note: This URL failed"):
@@ -4582,6 +4584,99 @@ def _strip_tool_tags(text: str) -> str:
     return "\n".join(out).strip()
 
 
+def _md_scalar(v) -> str:
+    if v is None or v == "" or v == [] or v == {}:
+        return "none"
+    if isinstance(v, bool):
+        return "yes" if v else "no"
+    return str(v)
+
+
+def _md_lines(value, depth=0):
+    """Nested dict/list -> markdown bullet lines (no raw braces, quotes or brackets)."""
+    pad = "  " * depth
+    lines = []
+    if isinstance(value, dict):
+        for k, v in value.items():
+            label = str(k).replace("_", " ").strip()
+            label = label[:1].upper() + label[1:]
+            if isinstance(v, (dict, list, tuple)) and v:
+                lines.append(f"{pad}- **{label}:**")
+                lines.extend(_md_lines(v, depth + 1))
+            else:
+                lines.append(f"{pad}- **{label}:** {_md_scalar(v)}")
+    else:
+        for item in value:
+            if isinstance(item, (dict, list, tuple)) and item:
+                lines.extend(_md_lines(item, depth + 1))
+            else:
+                lines.append(f"{pad}- {_md_scalar(item)}")
+    return lines
+
+
+def _fallback_recap(steps, last_text="") -> str:
+    """Task #131: the model ended its turn with no final text (e.g. only reasoning). Build a
+    short readable recap from the steps that did run instead of showing '[No result]'."""
+    tried = []
+    for st in (steps or [])[-8:]:
+        did = str(st.get("did") or "").strip().replace(chr(10), " ")
+        m = re.search(r"([a-z_]+\([^\n]{0,70})", did)
+        label = (m.group(1) if m else did)[:70].strip()
+        if label:
+            tried.append("- " + label)
+    last = str(last_text or "").strip()
+    if not tried and not last:
+        return ""
+    out = ["**Status:** the model finished without writing a summary."]
+    if tried:
+        out += ["", "**What was run**"] + tried
+    if last:
+        out += ["", "**Last result**", last[:600]]
+    return chr(10).join(out)
+
+
+def _humanize_answer(result) -> str:
+    """Task #131: a final answer that is a dict/list (or a string that is just a JSON /
+    Python-dict literal, as DeepSeek produces) is rendered as readable markdown instead of
+    being shown as its raw repr. Plain prose is returned untouched."""
+    if result is None:
+        return ""
+    data = result
+    if isinstance(result, str):
+        s = result.strip()
+        if len(s) < 2 or s[0] not in "{[" or s[-1] not in "}]":
+            return result
+        data = None
+        try:
+            data = json.loads(s)
+        except Exception:  # noqa: BLE001
+            try:
+                import ast
+                data = ast.literal_eval(s)
+            except Exception:  # noqa: BLE001
+                return result
+        if not isinstance(data, (dict, list)) or not data:
+            return result
+    elif isinstance(result, tuple):
+        data = list(result)
+    elif not isinstance(result, (dict, list)):
+        return str(result)
+    if isinstance(data, dict):
+        out = []
+        for k, v in data.items():
+            label = str(k).replace("_", " ").strip()
+            label = label[:1].upper() + label[1:]
+            if isinstance(v, (dict, list, tuple)) and v:
+                out.append("")
+                out.append(f"**{label}**")
+                out.extend(_md_lines(v, 0))
+            else:
+                out.append("")
+                out.append(f"**{label}:** {_md_scalar(v)}")
+        return chr(10).join(out).strip()
+    return chr(10).join(_md_lines(data, 0)).strip()
+
+
 _THINK_BLOCK = re.compile(r"<think>.*?</think>", re.DOTALL | re.IGNORECASE)
 
 # Malformed variants of smolagents' code-block tags. Weak coding models routinely
@@ -4812,6 +4907,62 @@ def _dsml_to_code_block(text):
     return (prose + "\n" if prose else "") + "<code>\n" + body + "\n</code>"
 
 
+# DeepSeek plain tool-call markup (issue #133). Besides the DSML form above, DeepSeek-family
+# models answer with <tool_calls><invoke name="X"><parameter name="k" string="true">v</parameter>
+# </invoke></tool_calls>. That matched nothing, so it reached the user as the FINAL answer (an
+# unexecuted run_command and no recap). Read it as a code block so the call runs, and strip any
+# leftover from the final text so it is never shown.
+_TC_INVOKE = re.compile(r"<invoke\s+name\s*=\s*[\x22\x27]([A-Za-z_]\w*)[\x22\x27]\s*>([\s\S]*?)(?:</invoke\s*>|(?=<invoke\b)|$)")
+_TC_PARAM = re.compile(r"<parameter\s+name\s*=\s*[\x22\x27]([A-Za-z_]\w*)[\x22\x27]"
+                       r"(?:\s+string\s*=\s*[\x22\x27](true|false)[\x22\x27])?\s*>([\s\S]*?)(?:</parameter\s*>|(?=<parameter\b)|$)")
+_TC_BLOCK = re.compile(r"<\s*tool_calls\s*>[\s\S]*?(?:<\s*/\s*tool_calls\s*>|$)", re.I)
+_TC_STRAY = re.compile(r"</?\s*tool_calls\s*>|</?\s*invoke\b[^<>]*>|</?\s*parameter\b[^<>]*>", re.I)
+_TC_JSON_WORDS = {"true": "True", "false": "False", "null": "None"}
+
+
+def _strip_toolcall_markup(text):
+    """Remove tool_calls/invoke/parameter markup from text that is shown to the user."""
+    if not text or ("<tool_calls" not in text and "<invoke" not in text and "<parameter" not in text):
+        return text
+    return _TC_STRAY.sub("", _TC_BLOCK.sub("", text)).strip()
+
+
+def _toolcalls_to_code_block(text):
+    """Rewrite tool_calls/invoke/parameter markup into a code block that calls the tools, or
+    return the text unchanged (only when there is no code tag already and the result COMPILES)."""
+    if not text or "<invoke" not in text or re.search(r"<\s*/?\s*code\b", text, re.I):
+        return text
+    calls = []
+    for m in _TC_INVOKE.finditer(text):
+        args = []
+        for pm in _TC_PARAM.finditer(m.group(2)):
+            val = pm.group(3).strip("\n")
+            lit = repr(val)
+            if pm.group(2) == "false":
+                if val in _TC_JSON_WORDS:
+                    lit = _TC_JSON_WORDS[val]
+                else:
+                    try:
+                        json.loads(val)
+                        lit = val
+                    except Exception:  # noqa: BLE001
+                        lit = repr(val)
+            args.append(pm.group(1) + "=" + lit)
+        calls.append("print(" + m.group(1) + "(" + ", ".join(args) + "))")
+    if not calls:
+        return text
+    body = "\n".join(calls)
+    try:
+        compile(body, "<gu-code>", "exec")
+    except Exception:  # noqa: BLE001
+        return text
+    first = text.find("<tool_calls")
+    if first < 0:
+        first = text.find("<invoke")
+    prose = text[:first].rstrip()
+    return (prose + "\n" if prose else "") + "<code>\n" + body + "\n</code>"
+
+
 def _code_block_readings(text):
     """Every plausible reading of where the code block is, CHEAPEST FIRST.
 
@@ -4876,6 +5027,10 @@ def _canonicalise_code_reply(text):
     if _dsml != text:
         _log("DSML tool-call markup converted to a code block (#119)")
         return _dsml
+    _tc = _toolcalls_to_code_block(text)
+    if _tc != text:
+        _log("tool_calls/invoke markup converted to a code block (#133)")
+        return _tc
     if not text or not _CODE_ANY_TAGGISH.search(text):
         return text
     fallback = None
@@ -9538,7 +9693,8 @@ def run_agent_chat(cfg):
         "instead of stopping the block. Keep UNRELATED changes in separate blocks. Take a separate "
         "step only when a call depends on a result you have not seen yet (e.g. web_search to "
         "find a URL, THEN fetch_url to read it). You SEE the Observation before the next step. "
-        "When you have enough to answer, call final_answer(<your answer>) in a code block. "
+        "When you have enough to answer, call final_answer(<your answer>) in a code block — pass a readable "
+        "markdown/prose recap (what you did, what you verified, what is left), never a dict, list or JSON. "
         "Prefer FEWER steps, and never repeat a tool call that already succeeded.\n\n"
         "{{TOOL_LIST}}\n"  # filled in below, once from the REAL registered tool objects
         + _skills_mod.prompt_block(_skills)
@@ -14227,7 +14383,9 @@ def run_agent_chat(cfg):
             _emit({"type": "step", "text": _retr})
             _log(f"retrieval summary: {_c(_retr)}")
         _emit({"type": "step", "text": "Composing answer…"})
-        final_text = _strip_tool_tags(str(result).strip()) or "[No result]"
+        final_text = (_strip_tool_tags(_humanize_answer(result).strip())
+                      or _fallback_recap(_turn_steps, _last_obs.get("text"))
+                      or "[No result]")
         _log(f"done (deterministic, no summarizer call): {len(final_text)} chars")
         _emit({"type": "final", "text": final_text})
         # onTurnEnd (task #270): after the answer is out, so a slow hook never delays it.

package/gu-cli.mjs

@@ -11,7 +11,7 @@
  */
 // FIRST, before anything that could itself fail on an old runtime: npm only WARNS about
 // `engines`, so an unsupported Node installs fine and then breaks somewhere unrelated.
-import { requireSupportedNode } from "./node-version.mjs";
+import { localModeNodeProblem, requireSupportedNode } from "./node-version.mjs";
 // The brand accent, shared with the REPL — see accent.mjs.
 import { accent } from "./accent.mjs";
 import { paint } from "./theme.mjs";
@@ -1068,13 +1068,16 @@ if (args[0] === "fix-path") {
     console.log("  If `gu` is not found in PowerShell: gu-cli fix-powershell");
     process.exit(0);
   }
-  const { globalPrefixOf } = await import("./install-path.mjs");
+  const { fixPathPrefix } = await import("./install-path.mjs");
   const { fileURLToPath } = await import("node:url");
   const { dirname, join } = await import("node:path");
   const here = dirname(fileURLToPath(import.meta.url));
-  const prefix = globalPrefixOf(here);
+  // #94: under `npx @tiens.nguyen/gu-cli fix-path` this file runs from npx's cache; fixPathPrefix
+  // then falls back to npm's global prefix when gu is installed there.
+  const prefix = fixPathPrefix({ here });
   if (!prefix) {
-    console.log(`  This gu is not a global npm install (${here}) — there is no PATH to fix.`);
+    console.log(`  gu is not installed globally with npm (${here}) — there is no PATH to fix.`);
+    console.log("  Install it with: npm install -g @tiens.nguyen/gu-cli");
     process.exit(0);
   }
   const binDir = join(prefix, "bin");
@@ -1089,9 +1092,24 @@ if (args[0] === "fix-path") {
     env: { ...env, npm_config_prefix: prefix, npm_config_global: "true" },
   });
   const said = `${r.stdout ?? ""}${r.stderr ?? ""}`.trim();
-  // It prints only when it wrote something. Silence means the line is already in the profile
-  // and this terminal simply predates it.
-  console.log(said || `  Your shell profile already adds ${binDir} to PATH — open a new terminal.`);
+  // It prints only when it wrote something. Silence USUALLY means the line is already in the
+  // profile — but check (#94): it is also silent when it skipped (e.g. a temp-dir prefix), and
+  // saying "already adds" then is a lie the user acts on.
+  if (said) {
+    console.log(said);
+  } else {
+    const { existsSync, readFileSync } = await import("node:fs");
+    const { homedir } = await import("node:os");
+    const has = [".zshrc", ".bashrc", ".bash_profile", ".profile"].some((f) => {
+      const p = join(homedir(), f);
+      try { return existsSync(p) && readFileSync(p, "utf8").includes("# >>> gu-cli PATH >>>"); } catch { return false; }
+    });
+    if (has) console.log(`  Your shell profile already adds ${binDir} to PATH — open a new terminal.`);
+    else {
+      console.log(`  ${binDir} is not on your PATH and was not added to a shell profile. Add this line to it:`);
+      console.log(`  export PATH="${binDir}:$PATH"`);
+    }
+  }
   process.exit(r.status ?? 0);
 }
 
@@ -1285,14 +1303,24 @@ if (args[0] === "clear" || args[0] === "reset") {
   }
   // The DATABASE first, while ~/.gonext/api still exists — dropping it needs the mongodb
   // driver that ships inside the downloaded API, and clearState is about to delete that folder.
-  if (groups.includes("work") && dbItem) {
+  // GitHub #125: the API's mongodb 7 driver needs Node 20.19+. On Node 18 it died with a bare
+  // "crypto is not defined"; say what is wrong and what to do instead.
+  const dropNodeProblem = groups.includes("work") && dbItem ? localModeNodeProblem() : "";
+  if (dropNodeProblem) {
+    console.log(
+      "could not drop the local database — " + dropNodeProblem.split("\n")[0] +
+        "\n  Upgrade Node and run `gu clear` again to drop it; the next setup may inherit its settings."
+    );
+  } else if (groups.includes("work") && dbItem) {
+    // GitHub #126: the client is closed even when connect/drop throws — an open client keeps
+    // its sockets and timers alive.
+    let client = null;
     try {
       const ce = await import("./client-expert-api.mjs");
       const { MongoClient } = ce.loadMongoDriver(apiDirForDrop);
-      const client = new MongoClient(ce.localMongoUri(), { serverSelectionTimeoutMS: 8000 });
+      client = new MongoClient(ce.localMongoUri(), { serverSelectionTimeoutMS: 8000 });
       await client.connect();
       await client.db(ce.LOCAL_DB_NAME).dropDatabase();
-      await client.close();
       console.log("dropped the local database");
     } catch (e) {
       // Said out loud, not swallowed: a machine that keeps its old settings row will look
@@ -1300,10 +1328,18 @@ if (args[0] === "clear" || args[0] === "reset") {
       console.log(
         `could not drop the local database (${e.message}) — the next setup may inherit its settings`
       );
+    } finally {
+      if (client) {await client.close(true).catch(() => {});}
     }
   }
-  const removed = await cs.clearState(items);
-  console.log(`removed ${removed.length} item(s).`);
+  // GitHub #126: whatever clearState does, gu clear ends here and exits.
+  try {
+    const removed = await cs.clearState(items);
+    console.log(`removed ${removed.length} item(s).`);
+  } catch (e) {
+    console.log(`could not remove everything (${e.message})`);
+    process.exit(1);
+  }
   console.log(
     groups.includes("identity")
       ? "\nNext `gu` starts from the beginning — it will ask which mode to run.\n"

package/gu-repl.mjs

@@ -676,6 +676,8 @@ let clientExpertRanThisProcess = false;
 // Declared HERE, above the first-run block: that block runs at module top level, before any
 // declaration further down exists.
 let openPresetsAtStart = false;
+// #132 follow-up: a hosted provider picked in first-run setup; its key and model are asked once the terminal exists.
+let addProviderAtStart = null;
 
 if (!workerKey || !apiBase) {
   if (process.stdin.isTTY && process.stdout.isTTY) {
@@ -800,6 +802,7 @@ if (!workerKey || !apiBase) {
           const sp = setupPrompts(rl0, { style: accent });
           const setup = await runClientExpertSetup({ ask: ask0, askStep: sp.askStep, askSecret: sp.askSecret });
           if (setup.loadPreset) openPresetsAtStart = true;
+          if (setup.addProvider) addProviderAtStart = setup.addProvider;
           // STOP when setup cannot go further.
           //
           // This used to fall through to the hosted sign-in, on the theory that "this machine
@@ -5294,7 +5297,8 @@ async function markInterrupted(cwd, history, task, how) {
   const status = interruptStatus(how);
   if (!status) return;
   last.status = status;
-  await saveSession(cwd, history, { unfinishedTask: sessionUnfinishedTask || task });
+  const keep = sessionUnfinishedTask || task;
+  await saveSession(cwd, history, { unfinishedTask: isResumeWord(keep) ? "" : keep }); // #201
 }
 
 async function rememberUnfinishedTask(cwd, task) {
@@ -6028,8 +6032,8 @@ async function reconfigureOllamaServer() {
   // with a single row is a question with one answer.
   let chat = "";
   if (agent) {
-    const cands = os.agentCandidates(agent.models);
-    const agentDefault = os.defaultAgentModel(
+    // Task #36: the saved agent stays a choice (and the default) even when it is not a gemma4.
+    const { cands, current: agentDefault } = os.agentChoices(
       agent.models,
       String(settings.agentModelName ?? "").trim()
     );
@@ -6239,7 +6243,9 @@ async function reconfigureOllamaServer() {
   // qwen3-embedding-only box) was written to settings above, so REQUIRED_EMBED_MODEL would
   // falsely scream "changed" while the seeded value is exactly what settings now hold.
   const seededEmbed = agent ? (os.defaultEmbedModel?.(agent.models) ?? os.REQUIRED_EMBED_MODEL) : "";
-  if (agent && String(settings.ragEmbedModel ?? "") !== seededEmbed) {
+  // Only a CHANGE when there was an embedder before — from none it is simply set (task #36).
+  const prevEmbed = String(settings.ragEmbedModel ?? "").trim();
+  if (agent && prevEmbed && prevEmbed !== seededEmbed) {
     console.log(dim("   the embedder changed — existing RAG indexes need rebuilding (/rag)."));
   }
   console.log(dim("   takes effect on your next question.\n"));
@@ -6524,8 +6530,24 @@ async function changeProviderModel(ps, rows, p) {
 /** Point a provider at a different URL, keeping its id so saved picks still resolve. */
 async function changeProviderUrl(ps, rows, p) {
   const md = await import("./model-doctor.mjs");
-  console.log("\n" + GUTTER + dim(`Enter keeps ${p.url}`));
-  const typed = (await ask("   provider URL: ")).trim();
+  // THE SAME PRESET LIST AS ADDING (#132), OpenAI-compatible ones only: an Ollama row has another
+  // kind, and this flow keeps the row's id and kind. "type a URL…" is the default row, so Enter
+  // still gets the old free-text question.
+  const urlPresets = ps.providerPresets("coding").filter((x) => (x.dialect ?? "openai") === "openai");
+  const urlRows = [...urlPresets.map((x) => `${x.name}${dim(`  · ${x.url}`)}`), "type a URL…"];
+  console.log("\n" + GUTTER + dim(`Which provider?  (${PICK_HINT})`));
+  const upi = await pickFromList(urlRows, urlRows.length - 1);
+  if (upi < 0) {
+    console.log(GUTTER + dim("(nothing changed)\n"));
+    return;
+  }
+  let typed;
+  if (upi < urlPresets.length) {
+    typed = urlPresets[upi].url;
+  } else {
+    console.log(GUTTER + dim(`Enter keeps ${p.url}`));
+    typed = (await ask("   provider URL: ")).trim();
+  }
   if (!typed || typed === p.url) {
     console.log(GUTTER + dim("(nothing changed)\n"));
     return;
@@ -6679,13 +6701,14 @@ async function addServer({ role = "coding", dialect = "openai", url: presetUrl =
     // without Claude, which serves no embeddings).
     const presets = ps.providerPresets(role);
     const rowsShown = [...presets.map((p) => `${p.name}${dim(`  · ${p.url}`)}`), "another URL…"];
+    // A preset names its own dialect (Ollama cloud is "ollama"); the rest are OpenAI-compatible.
     console.log(dim(`   Which provider?  (${PICK_HINT})`));
     const pi = await pickFromList(rowsShown, 0);
     if (pi < 0) {
       console.log(dim("   (nothing saved)\n"));
       return;
     }
-    if (pi < presets.length) target = as.serverTarget("openai", presets[pi].url);
+    if (pi < presets.length) target = as.serverTarget(presets[pi].dialect ?? "openai", presets[pi].url);
   }
   if (!target && !presetUrl) {
     const dflt = as.defaultServerUrl(role, dialect, settings, existing);
@@ -6861,7 +6884,8 @@ async function addServer({ role = "coding", dialect = "openai", url: presetUrl =
     rowId = ps.mintProviderId(existing.map((b) => String(b.id ?? "")));
     const { row } = as.rowForSave(role, { id: rowId, kind: target.dialect === "ollama" ? "ollama" : "openai", url: target.url, model, extras });
     rows = [...existing, row];
-  } else {
+  } else if (found) {
+    // A keyless agent/embed server with no row needs no row: the role's own setting holds its URL.
     const r = as.rowForSave(role, { existing: found.row, model });
     if (r.full) {
       console.log(yellow(`   ${host} already offers ${ps.MAX_MODELS_PER_SERVER} models — remove one in /model → manage providers first.\n`));
@@ -7102,6 +7126,14 @@ async function chooseModel() {
   const chosen = choices[idx];
   // Empty override means "use the account default" — so don't send an override when the
   // user picks the default (keeps the payload clean and lets a later default change win).
+  // GitHub #130: what to go back to if the key prompt for the new coder is cancelled (Esc).
+  const prevCoder = {
+    model: sessionCodingModel,
+    label: sessionCodingLabel,
+    url: sessionCodingUrl,
+    where: activeCoderWhere,
+    tail: sessionApiKeyTail,
+  };
   sessionCodingModel = chosen.id === def ? "" : chosen.id;
   setModelFallback(null); // #343: the warning named the coder that was just replaced
   // Remember the friendly name for the token footer, which should never show a raw
@@ -7143,7 +7175,23 @@ async function chooseModel() {
   // exactly this key (provider-scoped), so mirror it here; a keyless coder shows no tail.
   sessionApiKeyTail = ks.needsKey && !ks.missing && ks.stored ? ks.stored.slice(-4) : "";
   if (!ks.needsKey) keyChoicePending = false; // Ollama or local MLX — nothing to choose
-  if (ks.needsKey && ks.missing) await promptForCodingKey(sessionCodingModel);
+  if (ks.needsKey && ks.missing) {
+    try {
+      await promptForCodingKey(sessionCodingModel);
+    } catch (e) {
+      if (!wasPromptCancelled(e)) throw e;
+      // GitHub #130: Esc at the new coder's key prompt cancels the PICK, not just the key —
+      // otherwise gu is left on a coder it cannot use. Restore and re-save the previous one.
+      sessionCodingModel = prevCoder.model;
+      sessionCodingLabel = prevCoder.label;
+      sessionCodingUrl = prevCoder.url;
+      activeCoderWhere = prevCoder.where;
+      sessionApiKeyTail = prevCoder.tail;
+      await saveSessionCoding(resolve(process.cwd()));
+      console.log(dim("   (cancelled — the coding model was not changed)\n"));
+      return;
+    }
+  }
   else if (ks.needsKey) {
     // THIS FOLDER'S KEY, not the account's (task #222). ks.stored comes from GET /coding-key
     // with no job named, which answers with the account default — so after picking "accounting"
@@ -10210,6 +10258,16 @@ let forgetActiveStatus = null;
 let toggleActiveThought = null;
 /** #324: the screen under the Yes/No picker when it opened — the list picker's bookkeeping (#321). */
 let approvalOpenedOver = null;
+// #135: deadline (epoch ms) of the auto-"No" for the Yes/No prompt and for a timed list picker, 0 = none.
+// The remaining time is drawn next to the option that will be chosen, and ticks once a second.
+let approvalDeadline = 0;
+let listDeadline = 0;
+let listTimeoutIndex = -1;
+/** "2:49" above a minute, "45s" below; never negative. */
+function countdownLabel(msLeft) {
+  const s = Math.max(0, Math.ceil(msLeft / 1000));
+  return s >= 60 ? Math.floor(s / 60) + ":" + String(s % 60).padStart(2, "0") : s + "s";
+}
 function drawApprovalOptions(redraw) {
   // Two option lines, redrawn in place. On redraw the cursor sits one row below "No";
   // step up 2 rows, clear+rewrite both, landing back where we started.
@@ -10218,16 +10276,19 @@ function drawApprovalOptions(redraw) {
   // Both rows are a handful of columns wide, so they survive any resize on ONE physical row
   // each and the \x1b[2A count stays exact — which is why a resize can simply redraw here.
   const yes = approvalSel === 0 ? accent("▸ Yes") : dim("  Yes");
-  const no = approvalSel === 1 ? accent("▸ No") : dim("  No");
+  const tail = approvalDeadline ? "  (auto-No in " + countdownLabel(approvalDeadline - Date.now()) + ")" : "";
+  const no = approvalSel === 1 ? accent("▸ No" + tail) : dim("  No" + tail);
+  const noRow = clipVisible(GUTTER + no, Math.max(20, process.stdout.columns || 80) - 1);
   if (!redraw && barPinned() && !scrolledBack()) {
     const content = displayRows(transcript.lines(), Math.max(20, process.stdout.columns || 80) - 1).length;
     approvalOpenedOver = { content, pad: currentPad(content), height: viewportHeight() };
   }
-  regionUi((redraw ? "\x1b[2A" : "") + "\x1b[K" + GUTTER + yes + "\n\x1b[K" + GUTTER + no + "\n");
+  regionUi((redraw ? "\x1b[2A" : "") + "\x1b[K" + GUTTER + yes + "\n\x1b[K" + noRow + "\n");
 }
-function finishApproval(allow) {
+function finishApproval(allow, timedOut = false) {
   if (!approvalActive) return;
   approvalActive = false;
+  approvalDeadline = 0;
   // #324: on a full screen the two option rows SCROLLED the conversation up by two, and the one
   // result line below takes only one of them back — the content now ends a row above the region's
   // bottom. Unrecorded, every row put back from the transcript afterwards (the slash menu, a
@@ -10238,7 +10299,7 @@ function finishApproval(allow) {
     setPad(blockClosedPad({ contentRows: over.content, pad: over.pad, height: over.height, blockRows: 2 }), over.content);
   }
   // Settle the picker: overwrite the two option lines with a single result line.
-  process.stdout.write("\x1b[2A\x1b[J" + GUTTER + (allow ? accent("▸ Yes — running it") : red("▸ No — skipped")) + "\n");
+  process.stdout.write("\x1b[2A\x1b[J" + GUTTER + (allow ? accent("▸ Yes — running it") : red(timedOut ? "▸ No — no answer in time, skipped" : "▸ No — skipped")) + "\n");
   const r = approvalResolve;
   approvalResolve = null;
   if (r) r(!!allow);
@@ -10395,11 +10456,17 @@ async function approvalPrompt(command) {
     approvalSel = 0;
     approvalResolve = resolve;
     approvalActive = true;
+    approvalDeadline = Date.now() + 170000;
     drawApprovalOptions(false);
-    const t = setTimeout(() => finishApproval(false), 170000);
+    const t = setTimeout(() => finishApproval(false, true), 170000);
+    // #135: redraw the two option rows once a second so the countdown on "No" ticks down.
+    const tick = setInterval(() => {
+      if (approvalActive) drawApprovalOptions(true);
+    }, 1000);
     const orig = resolve;
     approvalResolve = (v) => {
       clearTimeout(t);
+      clearInterval(tick);
       orig(v);
     };
   });
@@ -10464,7 +10531,8 @@ function drawList(redraw) {
     // and "this is where you are" — and collapsing them into one marker is what makes a
     // multi-select feel like it forgot what you chose.
     const box = listMulti ? (listChecked.has(i) ? "[x] " : "[ ] ") : "";
-    const body = box + listItems[i];
+    const left = listDeadline && i === listTimeoutIndex ? "  (auto in " + countdownLabel(listDeadline - Date.now()) + ")" : "";
+    const body = box + listItems[i] + left;
     return clipVisible(GUTTER + (i === listSel ? accent("▸ " + body) : dim("  " + body)), width - 1);
   };
   const marker = (text) => clipVisible(GUTTER + dim(text), width - 1);
@@ -10653,6 +10721,12 @@ function pickFromList(items, initialSel = 0, opts = {}) {
     listSel = Math.max(0, Math.min(initialSel, items.length - 1));
     listResolve = resolve;
     listPickerActive = true;
+    listDeadline = 0;
+    listTimeoutIndex = -1;
+    if (opts && opts.timeoutMs > 0 && (opts.timeoutIndex ?? -1) >= 0) {
+      listDeadline = Date.now() + opts.timeoutMs;
+      listTimeoutIndex = opts.timeoutIndex;
+    }
     drawList(false);
     // Optional auto-resolve after opts.timeoutMs (task #81 time-budget prompt): if the
     // user doesn't choose in time, finish with opts.timeoutIndex (default -1). Wrap
@@ -10662,9 +10736,18 @@ function pickFromList(items, initialSel = 0, opts = {}) {
         () => finishList(opts.timeoutIndex ?? -1),
         opts.timeoutMs
       );
+      // #135: tick the countdown on the row that will be auto-chosen.
+      const tick = listDeadline
+        ? setInterval(() => {
+            if (listPickerActive) drawList(true);
+          }, 1000)
+        : null;
       const orig = listResolve;
       listResolve = (v) => {
         clearTimeout(to);
+        if (tick) clearInterval(tick);
+        listDeadline = 0;
+        listTimeoutIndex = -1;
         orig(v);
       };
     }
@@ -10971,6 +11054,8 @@ const answerFrom = (s) =>
 
 // Word-wrap a plain (ANSI-free) string into at most `maxLines` lines of width ≤ `width`,
 // appending an ellipsis to the last line when text was dropped (task #96 expand view).
+const THOUGHT_PREVIEW_LINES = 3; // #134: dim lines always shown under the cut purple thought
+const THOUGHT_OPEN_LINES = 8; // …and after a click / Ctrl+T
 const wrapThought = (s, width, maxLines) => {
   const words = String(s || "").split(/\s+/).filter(Boolean);
   const lines = [];
@@ -10992,7 +11077,13 @@ const wrapThought = (s, width, maxLines) => {
   if (truncated && lines.length) {
     const last = lines[maxLines - 1] ?? lines[lines.length - 1];
     const room = Math.max(1, width - 1);
-    lines[lines.length - 1] = (last.length > room ? last.slice(0, room) : last).replace(/\s+$/, "") + "…";
+    let kept = last;
+    if (last.length > room) {
+      // Cut at a word boundary (#134) so the "…" never lands inside a half word like "lim…".
+      const sp = last.lastIndexOf(" ", room);
+      kept = sp >= Math.floor(room * 0.5) ? last.slice(0, sp) : last.slice(0, room);
+    }
+    lines[lines.length - 1] = kept.replace(/\s+$/, "") + "…";
   }
   return lines;
 };
@@ -11674,12 +11765,9 @@ async function runAgentTurn(history, { attachJobId = "", ownJob = false } = {})
     // to reveal). Line 1 (what's happening now): a live command > the model's current Thought >
     // "almost done" > "Thinking". When clickable we drive line 1 from the SAME fuller string, so
     // the grey expansion CONTINUES where line 1 was cut off instead of repeating the head.
-    const clickable = !!fullLiveThought && !runningCmd && fullLiveThought.length > max;
-    // #350: the click and Ctrl+T ask the SAME question the highlight does. They used to accept
-    // any thought at all, so a line that was already whole took the click and redrew for nothing.
-    liveExpandable = clickable;
+    const split = !!fullLiveThought && !runningCmd && fullLiveThought.length > max;
     let label, rest;
-    if (clickable) {
+    if (split) {
       // Split at a WORD boundary near the line-1 width so nothing is cut mid-word: the last
       // space at/before max-1 (unless that snaps back too far into the line — then hard-cut).
       // Line 1 shows the head; the grey expansion CONTINUES from the same point (no dup, no
@@ -11699,7 +11787,15 @@ async function runAgentTurn(history, { attachJobId = "", ownJob = false } = {})
     }
     // When clicked, unfold the remainder as up to 4 wrapped GREY lines BETWEEN the thought and
     // the playful-word line.
-    const expLines = thoughtExpanded && rest ? wrapThought(rest, max, 4) : [];
+    // #134: the rest of a cut thought is ALWAYS visible as THOUGHT_PREVIEW_LINES dim lines under
+    // the purple one (ending in "…" when more was dropped); a click / Ctrl+T opens it up to
+    // THOUGHT_OPEN_LINES. #350: the click and Ctrl+T ask the SAME question the highlight does —
+    // they only count when opening actually shows more rows than the preview already does.
+    const previewLines = rest ? wrapThought(rest, max, THOUGHT_PREVIEW_LINES) : [];
+    const openLines = rest ? wrapThought(rest, max, THOUGHT_OPEN_LINES) : [];
+    const clickable = split && previewLines.join("\n") !== openLines.join("\n");
+    liveExpandable = clickable;
+    const expLines = thoughtExpanded ? openLines : previewLines;
     // WHAT IS RUNNING, under the label (task #177). Capped at two rows and clipped, because
     // this block is redrawn by walking the caret over PHYSICAL rows — a wrapped row here is
     // the #109/#138 failure mode.
@@ -12806,6 +12902,9 @@ async function main() {
       if (chosen === "client-expert") {
         const { runClientExpertSetup } = await import("./client-expert-run.mjs");
         const setup = await runClientExpertSetup({ ask });
+        // GitHub #126: setup refused (too-old Node) and said why — stop here, do not carry on
+        // into a REPL that cannot work and leave gu running after its last line.
+        if (setup?.stoppedAt === "node") process.exit(1);
         if (setup?.complete) reloadWorkerEnv();
       }
     } else if (!st.chosenMode && !gate.available) {
@@ -12829,8 +12928,10 @@ async function main() {
         askStep: (q) => ask(q).catch(toSkipAll),
         askSecret: (q) => askSecret(q).catch(toSkipAll),
       });
+      if (setup?.stoppedAt === "node") process.exit(1); // GitHub #126, as above
       if (setup?.complete) reloadWorkerEnv();
       if (setup?.loadPreset) openPresetsAtStart = true;
+      if (setup?.addProvider) addProviderAtStart = setup.addProvider;
     }
   } catch {
     /* never block the REPL on the mode question — the banner still shows what is configured */
@@ -13654,7 +13755,7 @@ async function main() {
       // rl.prompt() still runs for the NON-frozen fallback path, where readline draws the
       // input line itself. With the zone frozen (#141) discardTypedInput has already
       // repainted it, and prompt()'s own _refreshLine lands on the same paintBar().
-      rl.prompt();
+      if (!_stdinClosed) rl.prompt(); // ERR_USE_AFTER_CLOSE: a menu may have closed readline
     }
   };
   process.on("SIGINT", onInterrupt);
@@ -13685,6 +13786,17 @@ async function main() {
       console.log(GUTTER + dim(`(could not open the presets: ${e?.message ?? e} — run /use-presets)`));
     }
   }
+  // #132 follow-up: the provider chosen in first-run setup — the URL (or the preset list when none was
+  // chosen), then the key, then the model, all by the same flow as "add another provider".
+  if (addProviderAtStart) {
+    const pick = addProviderAtStart;
+    addProviderAtStart = null;
+    try {
+      await addServer({ role: "coding", dialect: pick.dialect || "openai", ...(pick.url ? { url: pick.url } : {}) });
+    } catch (e) {
+      console.log(GUTTER + dim(`(could not add the provider: ${e?.message ?? e} — run /model → add another provider)`));
+    }
+  }
 
   for (;;) {
     // #307: a held question whose turn ended without printing anything (and was not a quick
@@ -13701,6 +13813,9 @@ async function main() {
     enableClickToPosition();
     announceCliUpdate(); // deferred if a 6h refresh landed mid-turn; harmless once shown
     announceScaffold(); // the startup scaffold line, if it arrived while a turn was running
+    // ERR_USE_AFTER_CLOSE (task #36): Escape/EOF inside a menu can close readline; prompt() on a
+    // closed interface throws from here with a Node stack trace. Input is over — leave the loop.
+    if (_stdinClosed) break;
     rl.prompt();
     // #302, found live: prompt() puts the cursor at column 0, and a stopped turn's question is
     // already in the line — so what was typed went IN FRONT of it and Ctrl+U erased nothing.
@@ -13724,6 +13839,10 @@ async function main() {
       // prompt() erases from the caret down, so the row is rewritten in place and the empty
       // submit becomes a visual no-op. Nothing to step back over if the line came from
       // _lineQueue (no keypress, no echo) — hence the `echoed` guard.
+      if (_stdinClosed) {
+        lineRaw = null; // readline closed while we waited — never prompt() it again
+        break;
+      }
       if (process.stdout.isTTY) {
         // Frozen input line: nothing was echoed, so there is nothing to undo — just re-render.
         if (barPinned()) {
@@ -14174,19 +14293,17 @@ async function main() {
     // actually succeeds — saveSession writes the now-empty value then — so a terminal killed
     // mid-retry still has something to resume next time.
     sessionUnfinishedTask = "";
-    const { task: taskLine, resumed, nothingToResume } = resumeTask(line, pendingResume);
+    const resumeResult = resumeTask(line, pendingResume);
+    const { resumed, nothingToResume } = resumeResult;
+    // Nothing pending: the bare word goes to the coding model AS TYPED, with the conversation
+    // history, so the model decides what "continue" means from context (no hardcoded reply).
+    // The #201 loop stays closed at the WRITE side: rememberUnfinishedTask, saveSession and
+    // markInterrupted never store a resume word, and unfinishedTaskFromSession rejects one.
+    const taskLine = nothingToResume ? String(line ?? "").trim() : resumeResult.task;
     // #302 part 3: "continue" after a lost connection FOLLOWS the job that may still be running,
     // rather than sending the question again. Anything else typed lets it go.
     const reattach = lostTurn && isResumeWord(line) && lostTurn.taskLine === taskLine ? lostTurn : null;
     lostTurn = null;
-    // "continue" with nothing behind it runs NOTHING (task #201). It used to be sent to the
-    // model as the literal word, which then got remembered as the unfinished task — so every
-    // later "continue" retried "continue". Worse than the loop: with no task the model invents
-    // one, and the reported run edited a file and ran migrations nobody asked for.
-    if (nothingToResume) {
-      console.log(GUTTER + dim("nothing to continue — ask a question first.\n"));
-      continue;
-    }
     if (resumed) {
       const shown = taskLine.length > 80 ? taskLine.slice(0, 80) + "…" : taskLine;
       console.log(GUTTER + dim(`↻ retrying: ${shown}\n`));

package/install-path.mjs

@@ -56,7 +56,7 @@ if (isEntryPoint) {
     main();
   } catch (e) {
     // A postinstall that throws uninstalls the package. Report, swallow, succeed.
-    console.log(`gu-cli: PATH setup skipped (${e.message}) — run \`gu-cli fix-path\` if \`gu\` is not found.`);
+    console.log(`gu-cli: PATH setup skipped (${e.message}) — run \`npx @tiens.nguyen/gu-cli fix-path\` if \`gu\` is not found.`);
   }
   process.exit(0);
 }
@@ -83,6 +83,27 @@ export function globalPrefixOf(packageDir) {
   return prefix.join("/") || "/";
 }
 
+/**
+ * The global prefix `gu-cli fix-path` should repair — or "" (issue #94).
+ *
+ * WHY A FALLBACK: when npm stops running the postinstall, `gu-cli` is not on PATH either, so the
+ * way out is `npx @tiens.nguyen/gu-cli fix-path`. npx runs a COPY from ~/.npm/_npx/<hash>/, which
+ * is not a global install, so globalPrefixOf(here) is "". npx (and npm exec) still pass npm's
+ * global prefix in the env; use it only when gu really is installed there.
+ * Lookup order: where this file lives → npm_config_global_prefix → npm_config_prefix.
+ */
+export function fixPathPrefix({ here = "", env = process.env, exists = existsSync } = {}) {
+  const own = globalPrefixOf(here);
+  if (own) return own;
+  for (const p of [env.npm_config_global_prefix, env.npm_config_prefix]) {
+    if (!p) continue;
+    const prefix = resolve(String(p));
+    if (prefix.split("/").includes("node_modules")) continue;       // a project, not global
+    if (exists(join(prefix, "lib", "node_modules", "@tiens.nguyen", "gu-cli"))) return prefix;
+  }
+  return "";
+}
+
 function main() {
   if (platform() === "win32") return;                 // registry PATH + PowerShell alias: not here
   if (process.env.CI) return;                          // no user shell to help in CI

package/model-setup-ui.mjs

@@ -103,6 +103,34 @@ export function modelSetupMenu({ presets = 0 } = {}) {
   ];
 }
 
+/**
+ * The provider question inside "Guided" (#132 follow-up): Ollama stays the Enter default, then every
+ * preset (OpenAI, Claude, Kimi, Qwen...), then "another URL". `presets` = providerPresets().
+ */
+export function providerChoiceMenu(presets = []) {
+  return [
+    "Which provider do you want to use?",
+    "  1  Ollama server — this machine or your own   (Enter)",
+    ...presets.map((p, i) => `  ${i + 2}  ${p.name}  · ${p.url}`),
+    `  ${presets.length + 2}  Another provider — type its URL`,
+  ];
+}
+
+/** An answer to providerChoiceMenu → {kind:"ollama"|"preset"|"url"|"skip", preset?}, or null to ask again. */
+export function parseProviderChoice(raw, presets = []) {
+  if (raw === SKIP_ALL) return { kind: "skip" };
+  const t = String(raw ?? "").trim().toLowerCase();
+  if (t === "" || t === "1" || t === "ollama") return { kind: "ollama" };
+  if (t === "s" || t === "skip" || t === "n" || t === "no" || t === "-") return { kind: "skip" };
+  if (t === "u" || t === "url") return { kind: "url" };
+  if (/^\d+$/.test(t)) {
+    const n = Number(t);
+    if (n >= 2 && n <= presets.length + 1) return { kind: "preset", preset: presets[n - 2] };
+    if (n === presets.length + 2) return { kind: "url" };
+  }
+  return null;
+}
+
 /** An answer to the up-front question → "guided" | "preset" | "skip", or null to ask again. */
 export function parseModelSetupChoice(raw) {
   if (raw === SKIP_ALL) return "skip";

package/ollama-setup.mjs

@@ -93,6 +93,22 @@ export function defaultAgentModel(models, preferred = "") {
   )[0].model;
 }
 
+/**
+ * The AGENT choices /model offers (task #36): the gemma4 family, PLUS the agent already saved on
+ * this machine when this box has it. /agent-model may save any model (qwen3:14b, say); leaving it out
+ * here made the gemma4 default the preselected row, so Enter in /model silently replaced it.
+ * Returns { cands, current } — current is the row to preselect.
+ */
+export function agentChoices(models, saved = "") {
+  const list = Array.isArray(models) ? models : [];
+  const cands = agentCandidates(list);
+  const name = String(saved ?? "").trim();
+  const own = name && list.find((m) => m.model === name && !isLikelyEmbedder(m));
+  if (own && !cands.some((m) => m.model === name)) cands.unshift(own);
+  const current = own ? name : defaultAgentModel(list, name);
+  return { cands, current };
+}
+
 /** "nomic-embed-text" and "nomic-embed-text:latest" are the same model to a person. */
 const sameModel = (a, b) => {
   const bare = (s) => String(s ?? "").trim().toLowerCase().replace(/:latest$/, "");

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tiens.nguyen/gu-cli",
-  "version": "1.0.814",
+  "version": "1.0.817",
   "description": "Gu CLI — the gu terminal plus the local worker that runs agent / OCR / PDF / embedding jobs on your Mac (Ollama / MLX / OpenAI-compatible).",
   "type": "module",
   "license": "MIT",

package/proxy-dispatcher.mjs

@@ -14,6 +14,7 @@
  * when shouldProxy() says so.
  */
 import { shouldProxy, proxyUrl, redactSecrets } from "./proxy-config.mjs";
+import { nodeAtLeast } from "./node-version.mjs";
 
 /**
  * One ProxyAgent per dial string, kept for reuse.
@@ -25,6 +26,9 @@ import { shouldProxy, proxyUrl, redactSecrets } from "./proxy-config.mjs";
  */
 const agents = new Map();
 
+/** undici 7's engines floor (GitHub #127). */
+export const UNDICI_MIN_NODE = "20.18.1";
+
 let ProxyAgentCtor;
 let loadError = "";
 
@@ -38,6 +42,12 @@ let loadError = "";
  */
 async function proxyAgentCtor() {
   if (ProxyAgentCtor || loadError) return ProxyAgentCtor;
+  // GitHub #127: undici 7 needs Node 20.18.1+, while gu itself (Cloud mode) still runs on 18.
+  // Name the real cause instead of letting the import fail with an unrelated error.
+  if (!nodeAtLeast(process.version, UNDICI_MIN_NODE)) {
+    loadError = `the proxy needs Node ${UNDICI_MIN_NODE} or newer — this is ${process.version}`;
+    return ProxyAgentCtor;
+  }
   try {
     ({ ProxyAgent: ProxyAgentCtor } = await import("undici"));
   } catch (err) {

package/terminal-layout.mjs

@@ -1660,6 +1660,11 @@ export function classifyTurnFailure(message) {
   if (/The agent couldn't start|The agent's prompt (could not be loaded|could not be used)|agent's prompt templates could not be loaded/i.test(t)) {
     return "setup";
   }
+  // DEPS — the agent's Python framework is missing on this machine (GitHub #129). Every turn
+  // dies at import, before any model is called; neither retrying nor switching coder helps.
+  if (/No module named ['"]?(smolagents|openai)\b|ModuleNotFoundError:.*(smolagents|openai)/i.test(t)) {
+    return "deps";
+  }
   // CONFIG — the server is up and answering; it just has no model by that name (task #213).
   // Ollama says: model 'gemma4:e4b' not found  (and "…, try pulling it first" for embeddings).
   // Checked BEFORE the backend patterns because it arrives as a 404 from a perfectly healthy
@@ -1757,6 +1762,10 @@ export function retryAdvice({ resumable = true, failures = 1, kind = "task", min
       ? `gu's own machinery has failed ${failures} times in a row — not your question, and not the coding model. Check it before retrying: gu-cli api status, and gu-cli start if the worker is not running.`
       : 'that was gu\'s own machinery stopping, not your question — "continue" runs it again';
   }
+  if (kind === "deps") {
+    // GitHub #129: name the missing piece and the one step that fixes it.
+    return 'the agent framework (smolagents + openai) is not installed in this machine\'s Python, so nothing ran — run /doctor to install it, or restart gu and answer yes to "Install the agent framework now?", then "continue" sends your question again.';
+  }
   if (kind === "setup") {
     // Nothing ran, so there is nothing to diagnose in the model or the question — only the step the
     // message above names. "continue" is right AFTER that step, never before.