Update documentation links to point to the GitHub repository for various labs, enhancing accessibility and consistency across chapters. This includes updates in the README files for LLM fundamentals, prompt engineering, tools and function calling, RAG, agent architecture, state management, security and governance, observability, cost-latency engineering, data privacy, and production readiness index.

Author: kshvakov

book/01-llm-fundamentals/README.md

@@ -399,7 +399,7 @@ Not all models are equally good for agents.
 
 **Theory:** See [Appendix: Capability Benchmark](../appendix/README.md#capability-benchmark-characterization) — detailed description of what we test and why it's important.
 
-**Practice:** See [Lab 00: Model Capability Benchmark](../../labs/lab00-capability-check/README.md) — ready tool for testing the model.
+**Practice:** See [Lab 00: Model Capability Benchmark](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab00-capability-check) — ready tool for testing the model.
 
 ## Common Mistakes
 
@@ -492,7 +492,7 @@ Original history (2000 tokens):
 - **Trimming:** Quick one-time tasks, history not important
 - **Summarization:** Long sessions, contextual information important, autonomous agents
 
-See more: section "Context Optimization" in [Chapter 09: Agent Anatomy](../09-agent-architecture/README.md#context-optimization) and [Lab 09: Context Optimization](../../labs/lab09-context-optimization/README.md)
+See more: section "Context Optimization" in [Chapter 09: Agent Anatomy](../09-agent-architecture/README.md#context-optimization) and [Lab 09: Context Optimization](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab09-context-optimization)
 
 ### Mistake 3: LM Studio — Wrong Prompt Template (role support error)
 
@@ -546,7 +546,7 @@ If you don't know something, say "I don't know" or use a tool.`
 - [x] Know how to set `Temperature = 0` for deterministic behavior
 - [x] Understand context window limitations
 - [x] Know how to manage conversation history (summarization or trimming)
-- [x] Model supports Function Calling (verified via Lab 00)
+- [x] Model supports Function Calling (verified via [Lab 00](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab00-capability-check))
 - [x] System Prompt forbids hallucinations
 
 **Not completed:**

book/02-prompt-engineering/README.md

@@ -277,9 +277,9 @@ Assistant: {"tool": "check_status", "args": {"hostname": "web-01"}}  // Same for
 - **CoT → SOP:** SOP is a **process** fixed by instruction and/or examples. CoT helps the model follow this process step by step.
 
 See more:
-- **[Lab 01: Basics](../../labs/lab01-basics/README.md)** — working with context and memory
-- **[Lab 02: Tools](../../labs/lab02-tools/README.md)** — response format via Function Calling
-- **[Lab 06: Incident (SOP)](../../labs/lab06-incident/README.md)** — SOP as action algorithm
+- **[Lab 01: Basics](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab01-basics)** — working with context and memory
+- **[Lab 02: Tools](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab02-tools)** — response format via Function Calling
+- **[Lab 06: Incident (SOP)](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab06-incident)** — SOP as action algorithm
 
 ## End-to-End Example: What Exactly the Agent Sends to LLM
 

book/03-tools-and-function-calling/README.md

@@ -546,7 +546,7 @@ graph TB
 - The model sees the full context (system + user + tool call + tool result)
 - Generates final response: `"nginx is running normally, service ONLINE"`
 
-**Note:** For Go implementation examples, see [Lab 02: Tools](../../labs/lab02-tools/README.md) and [Lab 04: Autonomy](../../labs/lab04-autonomy/README.md)
+**Note:** For Go implementation examples, see [Lab 02: Tools](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab02-tools) and [Lab 04: Autonomy](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab04-autonomy)
 
 ### Key Points for Developers
 
@@ -570,7 +570,7 @@ graph TB
 
 See how to write instructions and examples: **[Chapter 02: Prompting](../02-prompt-engineering/README.md)**
 
-Practice: **[Lab 02: Tools](../../labs/lab02-tools/README.md)**, **[Lab 04: Autonomy](../../labs/lab04-autonomy/README.md)**
+Practice: **[Lab 02: Tools](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab02-tools)**, **[Lab 04: Autonomy](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab04-autonomy)**
 
 ### Why Is This Not Magic?
 

book/06-rag/README.md

@@ -168,7 +168,7 @@ msg2 := resp2.Choices[0].Message
 
 In this lab, we implement **simple RAG** (keyword search). In production, **vector search** (Semantic Search) is used, which searches by meaning, not by words.
 
-**Simple RAG (Lab 07):**
+**Simple RAG ([Lab 07](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab07-rag)):**
 ```go
 // Search by substring match
 if strings.Contains(content, query) {
@@ -696,7 +696,7 @@ var toolCatalog = []ToolDefinition{
 
 For tool search, you can use two approaches: simple keyword search (for learning) and vector search (for production).
 
-**Simple search (Lab 13):**
+**Simple search ([Lab 13](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab13-tool-retrieval)):**
 ```go
 func searchToolCatalog(query string, topK int) []ToolDefinition {
     // Simple search by description and tags

book/09-agent-architecture/README.md

@@ -315,7 +315,7 @@ Thought: I'll analyze results and formulate conclusions.
 
 ### Plan-and-Solve
 
-For complex tasks (Lab 06 Incident), ReAct may "get lost" in details.
+For complex tasks ([Lab 06](https://github.com/kshvakov/ai-agent-course/tree/main/labs/lab06-incident) Incident), ReAct may "get lost" in details.
 
 **Architecture:**
 

book/11-state-management/README.md

@@ -250,7 +250,7 @@ func resumeTask(taskID string) error {
 
 ### Integration Point 1: Agent Loop
 
-In `labs/lab04-autonomy/main.go` add state persistence:
+In [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go) add state persistence:
 
 ```go
 // At start of agent run:
@@ -275,7 +275,7 @@ saveTask(task)
 
 ### Integration Point 2: Tool Execution
 
-In `labs/lab02-tools/main.go` add retry for tools:
+In [`labs/lab02-tools/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab02-tools/main.go) add retry for tools:
 
 ```go
 func executeToolWithRetry(toolCall openai.ToolCall) (string, error) {
@@ -291,7 +291,7 @@ func executeToolWithRetry(toolCall openai.ToolCall) (string, error) {
 
 ## Mini Code Example
 
-Complete example with workflow and state management based on `labs/lab04-autonomy/main.go`:
+Complete example with workflow and state management based on [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go):
 
 ```go
 package main

book/17-security-and-governance/README.md

@@ -322,7 +322,7 @@ func logAudit(log AuditLog) {
 
 ### Integration Point 1: Tool Execution
 
-In `labs/lab02-tools/main.go` add access check and confirmation:
+In [`labs/lab02-tools/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab02-tools/main.go) add access check and confirmation:
 
 ```go
 func executeTool(toolCall openai.ToolCall, userRole UserRole) (string, error) {
@@ -357,7 +357,7 @@ func executeTool(toolCall openai.ToolCall, userRole UserRole) (string, error) {
 
 ### Integration Point 2: Human-in-the-Loop
 
-In `labs/lab05-human-interaction/main.go` confirmations already exist. Extend them for risk scoring:
+In [`labs/lab05-human-interaction/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab05-human-interaction/main.go) confirmations already exist. Extend them for risk scoring:
 
 ```go
 func requestConfirmation(toolCall openai.ToolCall) bool {
@@ -375,7 +375,7 @@ func requestConfirmation(toolCall openai.ToolCall) bool {
 
 ## Mini Code Example
 
-Complete example with security based on `labs/lab05-human-interaction/main.go`:
+Complete example with security based on [`labs/lab05-human-interaction/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab05-human-interaction/main.go):
 
 ```go
 package main

book/19-observability-and-tracing/README.md

@@ -94,7 +94,7 @@ func generateRunID() string {
 
 ### Step 3: Logging in Agent Loop
 
-Insert logging into the agent loop (see `labs/lab04-autonomy/main.go`):
+Insert logging into the agent loop (see [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go)):
 
 ```go
 func runAgent(ctx context.Context, client *openai.Client, userInput string) (string, error) {
@@ -212,7 +212,7 @@ func logAgentRunError(runID string, err error) {
 
 ### Step 5: Tool Execution Logging
 
-In tool execution functions (see `labs/lab02-tools/main.go`) add logging:
+In tool execution functions (see [`labs/lab02-tools/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab02-tools/main.go)) add logging:
 
 ```go
 func executeToolWithLogging(runID string, toolCall openai.ToolCall) (string, error) {
@@ -252,7 +252,7 @@ func executeToolWithLogging(runID string, toolCall openai.ToolCall) (string, err
 
 ### Integration Point 1: Agent Loop
 
-In `labs/lab04-autonomy/main.go` add logging to the agent loop:
+In [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go) add logging to the agent loop:
 
 ```go
 // At start of main():
@@ -271,7 +271,7 @@ log.Printf("TOOL_EXECUTED: run_id=%s tool=%s result=%s", runID, toolCall.Functio
 
 ### Integration Point 2: Tool Execution
 
-In `labs/lab02-tools/main.go` add logging when executing tools:
+In [`labs/lab02-tools/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab02-tools/main.go) add logging when executing tools:
 
 ```go
 // In tool execution function:
@@ -285,7 +285,7 @@ func executeTool(runID string, toolCall openai.ToolCall) (string, error) {
 
 ## Mini Code Example
 
-Complete example with observability based on `labs/lab04-autonomy/main.go`:
+Complete example with observability based on [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go):
 
 ```go
 package main
@@ -1033,7 +1033,7 @@ metrics.RecordLatency("tool_execution_duration", latency, map[string]string{"too
 
 ### Exercise 1: Implement Structured Logging
 
-Add structured logging to `labs/lab04-autonomy/main.go`:
+Add structured logging to [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go):
 
 ```go
 func logAgentRun(runID string, userInput string, toolCalls []ToolCall, result string) {

book/20-cost-latency-engineering/README.md

@@ -64,7 +64,7 @@ func checkTokenBudget(budget TokenBudget) error {
 
 ### Step 2: Iteration Limits
 
-Limit the number of ReAct loop iterations (see `labs/lab04-autonomy/main.go`):
+Limit the number of ReAct loop iterations (see [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go)):
 
 ```go
 func runAgent(ctx context.Context, client *openai.Client, userInput string) (string, error) {
@@ -155,7 +155,7 @@ func setCachedResult(key string, result string, ttl time.Duration) {
 
 ### Step 4: Model Routing by Complexity
 
-Use cheaper models for simple tasks (see `labs/lab09-context-optimization/main.go`):
+Use cheaper models for simple tasks (see [`labs/lab09-context-optimization/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab09-context-optimization/main.go)):
 
 ```go
 func selectModel(taskComplexity string) string {
@@ -240,7 +240,7 @@ func runAgentWithTimeout(ctx context.Context, client *openai.Client, userInput s
 
 ### Integration Point 1: Agent Loop
 
-In `labs/lab04-autonomy/main.go` add budget check and iteration limit:
+In [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go) add budget check and iteration limit:
 
 ```go
 const MaxIterations = 10
@@ -259,7 +259,7 @@ for i := 0; i < MaxIterations; i++ {
 
 ### Integration Point 2: Context Optimization
 
-In `labs/lab09-context-optimization/main.go` token counting already exists. Add budget check:
+In [`labs/lab09-context-optimization/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab09-context-optimization/main.go) token counting already exists. Add budget check:
 
 ```go
 func adaptiveContextManagement(ctx context.Context, client *openai.Client, messages []openai.ChatCompletionMessage, maxTokens int) []openai.ChatCompletionMessage {
@@ -276,7 +276,7 @@ func adaptiveContextManagement(ctx context.Context, client *openai.Client, messa
 
 ## Mini Code Example
 
-Complete example with cost control based on `labs/lab04-autonomy/main.go`:
+Complete example with cost control based on [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go):
 
 ```go
 package main
@@ -864,7 +864,7 @@ if err == context.DeadlineExceeded {
 
 ### Exercise 1: Implement Token Budget Check
 
-Add budget check to `labs/lab04-autonomy/main.go`:
+Add budget check to [`labs/lab04-autonomy/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab04-autonomy/main.go):
 
 ```go
 func checkTokenBudget(used int, limit int) error {

book/24-data-and-privacy/README.md

@@ -84,7 +84,7 @@ func logWithRedaction(runID string, data map[string]any) {
 
 ### Integration Point: User Input
 
-In `labs/lab05-human-interaction/main.go` sanitize input data:
+In [`labs/lab05-human-interaction/main.go`](https://github.com/kshvakov/ai-agent-course/blob/main/labs/lab05-human-interaction/main.go) sanitize input data:
 
 ```go
 userInput := sanitizePII(sanitizeSecrets(rawInput))