ngkore · chmodshubham · Jan 30, 2026 · Jan 30, 2026
diff --git a/.sphinx/_static/custom.css b/.sphinx/_static/custom.css
@@ -150,6 +150,21 @@ code.literal {
     line-height: 1.5;
     font-size: 0.9rem;
     overflow-x: auto;
+    white-space: pre !important;
+}
+
+/* Ensure code blocks preserve whitespace */
+pre, code.highlight pre, div.highlight pre {
+    white-space: pre !important;
+}
+
+/* Specific override for console/text code blocks */
+.highlight-console pre,
+.highlight-text pre,
+.highlight-none pre {
+    white-space: pre !important;
+    font-family: 'Courier New', Courier, monospace !important;
+    tab-size: 4;
 }
 
 /** Enhanced image styling **/

diff --git a/.sphinx/_static/ngkore-logo.png b/.sphinx/_static/ngkore-logo.png
diff --git a/ai-ml/ai-code-is-going-to-kill-your-startup.md b/ai-ml/ai-code-is-going-to-kill-your-startup.md
@@ -4,9 +4,8 @@
 
 **Published:** Nov 15, 2025
 
-*What happened when I watched the smartest security engineers in the world realize they don’t know what to do about AI-generated code*
+*What happened when I watched the smartest security engineers in the world realize they don't know what to do about AI-generated code*
 
----
 ![](./images/ai-code-openssl/openssl-conference.jpg)
 
 In a conference room in Prague sat some of the most experienced security engineers in the world — people who maintain OpenSSL, the people whose code encrypts your bank transactions, your medical records, and every password you’ve ever typed. They’ve seen every nightmare scenario: buffer overflows that crashed servers worldwide, Heartbleed, and timing attacks so subtle they took years to discover. These people have forgotten more about security than most of us will ever learn, and in 2025 they were genuinely unsure whether AI-generated code should ever touch their codebase.
@@ -250,5 +249,4 @@ You can ship features faster with AI. Prototype more quickly. Explore more ideas
 * Can LLMs Generate Correct and Secure Backends? (Vero et al., 2025): [arXiv 2502.11844](https://arxiv.org/abs/2502.11844)
 * Prompting Techniques for Secure Code Generation (Tony et al.): [arXiv 2407.07064](https://arxiv.org/abs/2407.07064)
 
----
 > *Now go fix your security practices before AI breaks them for you.*
diff --git a/ai-ml/devops/ai-automation-in-aws-with-mcp.md b/ai-ml/devops/ai-automation-in-aws-with-mcp.md
@@ -26,13 +26,13 @@ This architecture means you can give Qwen a high-level goal in natural language,
 
 The first step was to configure Qwen to use our three AWS-focused MCP servers. I started by checking the current setup.
 
-```shell
+```bash
 qwen mcp list
 ```
 
 As expected, it returned “No MCP servers configured.” I then applied a pre-configured settings file to connect Qwen to the servers.
 
-```shell
+```bash
 mkdir -p /root/.qwen
 cp /root/settings.json /root/.qwen/settings.json
 ```

diff --git a/ai-ml/devops/ai-incident-commander.md b/ai-ml/devops/ai-incident-commander.md
@@ -27,7 +27,7 @@ This structure allowed me to manage the incident strategically, focusing on the
 
 Before diving into the production fire, the first step was to ensure my AI team was online and ready. I ran a quick check to list the installed agents and verify the connection to our Plane ticketing system.
 
-```shell
+```bash
 # Verify available agents
 qwen --prompt "List installed agents" 2>/dev/null
 # Test Plane MCP Integration
@@ -42,7 +42,7 @@ The system confirmed three agents were active (cloud-architect, kubernetes-speci
 
 The most critical issue was the application downtime caused by the crashing pods. I navigated to the relevant directory and delegated the problem to our Kubernetes expert.
 
-```shell
+```bash
 cd /root/k8s-incident
 qwen -y 2>/dev/null
 ```
@@ -82,7 +82,7 @@ Think of Terraform code as the official architectural blueprint for a house. Dri
 
 This is precisely the problem our cloud-architect agent was designed to solve: to programmatically detect this drift, report on it, and bring our infrastructure back into alignment with our code.
 
-```shell
+```bash
 cd /root/terraform-static-site
 qwen -y 2>/dev/null
 ```

diff --git a/ai-ml/devops/building-virtual-devops-team-with-qwen-subagents.md b/ai-ml/devops/building-virtual-devops-team-with-qwen-subagents.md
@@ -31,7 +31,7 @@ This means we are not building a complex application or spinning up a new contai
 
 The first step was to “hire” my new team members by installing their agent files. A simple setup script handled the installation
 
-```shell
+```bash
 bash /root/setup-agents.sh
 ```
 
@@ -74,7 +74,7 @@ The new Dockerfile used a multi-stage build, switched to a slim Python base imag
 
 With the docker-optimized agent's work complete, it was time for the moment of truth. I built the new, optimized Docker image using the agent-generated Dockerfile.optimized
 
-```shell
+```bash
 docker build -f /root/production-issues/bad-docker/Dockerfile.optimized
 -t my-app:optimized /root/production-issues/bad-docker/
 Press enter or click to view image in full size
@@ -87,7 +87,7 @@ The build completed successfully, installing only the necessary dependencies and
 
 When I checked the final image size:
 
-```shell
+```bash
 docker images | grep my-app
 
 ```

diff --git a/ai-ml/k2-think-deployment-via-sglang.md b/ai-ml/k2-think-deployment-via-sglang.md
@@ -18,7 +18,7 @@ Let’s start with the prerequisites:
 
 - A K8s cluster installed with Nvidia GPU Operator
 
-```bash
+```console
 NAMESPACE         NAME                                                              READY   STATUS      RESTARTS      AGE
 cert-manager      cert-manager-5969544f77-pmrt7                                     1/1     Running     0             9d
 cert-manager      cert-manager-cainjector-65967ff5cc-tbntm                          1/1     Running     0             9d
@@ -113,7 +113,7 @@ prometheus        prometheus-prometheus-node-exporter-j4tjb
 
 - Storage class configured
 
-```bash
+```console
 kubectl get sc
 NAME                 PROVISIONER             RECLAIMPOLICY   VOLUMEBINDINGMODE      ALLOWVOLUMEEXPANSION   AGE
 local-path           rancher.io/local-path   Delete          WaitForFirstConsumer   true                   13d

diff --git a/ai-ml/openai-120b-model-deployment.md b/ai-ml/openai-120b-model-deployment.md
@@ -21,7 +21,7 @@ This guide will walk you through the deployment of OpenAI OSS Models
 
 Make sure you have a K8s cluster with NVidia GPU operator installed:
 
-```bash
+```console
 NAME                                                           READY   STATUS      RESTARTS       AGE
 gpu-feature-discovery-v2xpk                                    1/1     Running     0              4d8h
 gpu-operator-644fb64985-ffzws                                  1/1     Running     0              4d8h
@@ -381,4 +381,4 @@ spec:
 
 ### Reference
 
-- https://openai.com/index/introducing-gpt-oss/
+- <https://openai.com/index/introducing-gpt-oss/>