From 57f2c6485700e78d91a106b73123de4c96e736b1 Mon Sep 17 00:00:00 2001
From: latenighthackathon <latenighthackathon@users.noreply.github.com>
Date: Sun, 29 Mar 2026 14:52:25 -0500
Subject: [PATCH 1/2] fix(onboard): add timeout to spawnSync calls to prevent
 hung processes

All spawnSync calls in the onboarding wizard lacked a timeout option,
meaning a hung curl or stalled download would freeze the entire wizard
with no recovery path.

Add process-level timeouts as a safety net:
- 30s for curl endpoint probes (10s buffer over curl --max-time 20)
- 10min for ollama model downloads
- 5min for install-openshell.sh execution

Closes #1017
---
 bin/lib/onboard.js | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/bin/lib/onboard.js b/bin/lib/onboard.js
index 1480f435d..699e0ac87 100644
--- a/bin/lib/onboard.js
+++ b/bin/lib/onboard.js
@@ -678,6 +678,7 @@ function probeOpenAiLikeEndpoint(endpointUrl, model, apiKey) {
       const result = spawnSync("bash", ["-c", cmd], {
         cwd: ROOT,
         encoding: "utf8",
+        timeout: 30_000,
         env: {
           ...process.env,
           NEMOCLAW_PROBE_API_KEY: apiKey,
@@ -727,6 +728,7 @@ function probeAnthropicEndpoint(endpointUrl, model, apiKey) {
     const result = spawnSync("bash", ["-c", cmd], {
       cwd: ROOT,
       encoding: "utf8",
+      timeout: 30_000,
       env: {
         ...process.env,
         NEMOCLAW_PROBE_API_KEY: apiKey,
@@ -867,6 +869,7 @@ function fetchNvidiaEndpointModels(apiKey) {
     const result = spawnSync("bash", ["-c", cmd], {
       cwd: ROOT,
       encoding: "utf8",
+      timeout: 30_000,
       env: {
         ...process.env,
         NEMOCLAW_PROBE_API_KEY: apiKey,
@@ -920,6 +923,7 @@ function fetchOpenAiLikeModels(endpointUrl, apiKey) {
     const result = spawnSync("bash", ["-c", cmd], {
       cwd: ROOT,
       encoding: "utf8",
+      timeout: 30_000,
       env: {
         ...process.env,
         NEMOCLAW_PROBE_API_KEY: apiKey,
@@ -957,6 +961,7 @@ function fetchAnthropicModels(endpointUrl, apiKey) {
     const result = spawnSync("bash", ["-c", cmd], {
       cwd: ROOT,
       encoding: "utf8",
+      timeout: 30_000,
       env: {
         ...process.env,
         NEMOCLAW_PROBE_API_KEY: apiKey,
@@ -1223,6 +1228,7 @@ function pullOllamaModel(model) {
     cwd: ROOT,
     encoding: "utf8",
     stdio: "inherit",
+    timeout: 600_000,
     env: { ...process.env },
   });
   return result.status === 0;
@@ -1358,6 +1364,7 @@ function installOpenshell() {
     env: process.env,
     stdio: ["ignore", "pipe", "pipe"],
     encoding: "utf-8",
+    timeout: 300_000,
   });
   if (result.status !== 0) {
     const output = `${result.stdout || ""}${result.stderr || ""}`.trim();

From f3460d7530499ba60dd71154424468a3dc3b4009 Mon Sep 17 00:00:00 2001
From: latenighthackathon <latenighthackathon@users.noreply.github.com>
Date: Sun, 29 Mar 2026 15:05:15 -0500
Subject: [PATCH 2/2] fix(onboard): distinguish timeout from other failures in
 ollama pull

When spawnSync kills the process due to timeout, result.signal is
SIGTERM. Surface a specific error message so users know the 10-minute
limit was hit, rather than seeing the generic pull failure message.

Addresses CodeRabbit review nitpick.
---
 bin/lib/onboard.js | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/bin/lib/onboard.js b/bin/lib/onboard.js
index 699e0ac87..525d129f3 100644
--- a/bin/lib/onboard.js
+++ b/bin/lib/onboard.js
@@ -1231,6 +1231,10 @@ function pullOllamaModel(model) {
     timeout: 600_000,
     env: { ...process.env },
   });
+  if (result.signal === "SIGTERM") {
+    console.error(`  Model pull timed out after 10 minutes. Try a smaller model or check your network connection.`);
+    return false;
+  }
   return result.status === 0;
 }