htmlgraph 0.24.1__py3-none-any.whl → 0.25.0__py3-none-any.whl

htmlgraph/cli.py

@@ -41,6 +41,13 @@ Analytics:
     htmlgraph analytics                           # Project-wide analytics
     htmlgraph analytics --session-id SESSION_ID   # Single session analysis
     htmlgraph analytics --recent N                # Analyze recent N sessions
+
+Cost Attribution:
+    htmlgraph cigs cost-dashboard                 # Display cost summary in console
+    htmlgraph cigs cost-dashboard --save          # Save to .htmlgraph/cost-dashboard.html
+    htmlgraph cigs cost-dashboard --open          # Open in browser after generation
+    htmlgraph cigs cost-dashboard --json          # Output JSON instead of HTML
+    htmlgraph cigs cost-dashboard --output PATH   # Custom output path
 """
 
 import argparse
@@ -51,6 +58,20 @@ from datetime import datetime
 from pathlib import Path
 from typing import Any
 
+from rich import box
+from rich.console import Console
+from rich.panel import Panel
+from rich.progress import Progress, SpinnerColumn, TextColumn
+from rich.prompt import Confirm, Prompt
+from rich.table import Table
+from rich.traceback import install as install_traceback
+
+# Install Rich traceback globally for better error display
+install_traceback(show_locals=True)
+
+# Global Rich Console for beautiful CLI output
+console = Console()
+
 
 def create_json_response(
     command: str,
@@ -84,14 +105,15 @@ def cmd_install_gemini_extension(args: argparse.Namespace) -> None:
     extension_dir = package_dir / "extensions" / "gemini"
 
     if not extension_dir.exists():
-        print(f"Error: Gemini extension not found at {extension_dir}", file=sys.stderr)
-        print(
-            "The extension may not be bundled with this version of htmlgraph.",
-            file=sys.stderr,
+        console.print(
+            f"[red]Error: Gemini extension not found at {extension_dir}[/red]"
+        )
+        console.print(
+            "[red]The extension may not be bundled with this version of htmlgraph.[/red]"
         )
         sys.exit(1)
 
-    print(f"Installing Gemini extension from: {extension_dir}")
+    console.print(f"[cyan]Installing Gemini extension from:[/cyan] {extension_dir}")
 
     # Run gemini extensions install with the bundled path
     try:
@@ -101,35 +123,125 @@ def cmd_install_gemini_extension(args: argparse.Namespace) -> None:
             text=True,
             check=True,
         )
-        print(result.stdout)
-        print("\n✅ Gemini extension installed successfully!")
-        print("\nTo verify installation:")
-        print("  gemini extensions list")
+        console.print(result.stdout)
+        console.print("\n[green]✅ Gemini extension installed successfully![/green]")
+        console.print("\nTo verify installation:")
+        console.print("  gemini extensions list", style="dim")
     except subprocess.CalledProcessError as e:
-        print(f"Error installing extension: {e.stderr}", file=sys.stderr)
+        console.print(f"[red]Error installing extension: {e.stderr}[/red]", style="red")
         sys.exit(1)
     except FileNotFoundError:
-        print("Error: 'gemini' command not found.", file=sys.stderr)
-        print("Please install Gemini CLI first:", file=sys.stderr)
-        print("  npm install -g @google/gemini-cli", file=sys.stderr)
+        console.print("[red]Error: 'gemini' command not found.[/red]")
+        console.print("Please install Gemini CLI first:")
+        console.print("  npm install -g @google/gemini-cli", style="dim")
         sys.exit(1)
 
 
 def cmd_serve(args: argparse.Namespace) -> None:
-    """Start the HtmlGraph server."""
-    from htmlgraph.operations import start_server
+    """Start the HtmlGraph server (FastAPI-based)."""
+    import asyncio
 
-    start_server(
-        port=args.port,
-        graph_dir=args.graph_dir,
-        static_dir=args.static_dir,
-        host=args.host,
-        watch=not args.no_watch,
-        auto_port=args.auto_port,
+    from htmlgraph.operations.fastapi_server import (
+        run_fastapi_server,
+        start_fastapi_server,
     )
 
-    # The start_server operation already handles all output and blocks
-    # No additional CLI formatting needed
+    try:
+        # Default to database in graph dir if not specified
+        db_path = getattr(args, "db", None)
+        if not db_path:
+            db_path = str(Path(args.graph_dir) / "index.sqlite")
+
+        result = start_fastapi_server(
+            port=args.port,
+            host=args.host,
+            db_path=db_path,
+            auto_port=args.auto_port,
+            reload=getattr(args, "reload", False),
+        )
+
+        # Print server info
+        console.print("\n[bold cyan]HtmlGraph Server (FastAPI)[/bold cyan]")
+        console.print(f"URL: [bold blue]{result.handle.url}[/bold blue]")
+        console.print(f"Graph directory: {args.graph_dir}")
+        console.print(f"Database: {result.config_used['db_path']}")
+
+        if result.warnings:
+            for warning in result.warnings:
+                console.print(f"[yellow]Warning: {warning}[/yellow]")
+
+        from htmlgraph.server import HtmlGraphAPIHandler
+
+        console.print(f"Collections: {', '.join(HtmlGraphAPIHandler.COLLECTIONS)}")
+        console.print("\n[cyan]Features:[/cyan]")
+        console.print("  • Real-time agent activity feed (HTMX)")
+        console.print("  • Orchestration chains visualization")
+        console.print("  • Feature tracker with Kanban view")
+        console.print("  • Session metrics & performance analytics")
+        console.print("\n[cyan]Press Ctrl+C to stop.[/cyan]\n")
+
+        # Run server
+        asyncio.run(run_fastapi_server(result.handle))
+
+    except KeyboardInterrupt:
+        console.print("\n[yellow]Shutting down...[/yellow]")
+    except Exception as e:
+        console.print(f"\n[red]Error:[/red] {e}")
+        import traceback
+
+        if getattr(args, "verbose", False):
+            traceback.print_exc()
+        sys.exit(1)
+
+
+def cmd_serve_api(args: argparse.Namespace) -> None:
+    """Start the FastAPI-based observability dashboard (Phase 3)."""
+    import asyncio
+
+    from htmlgraph.operations.fastapi_server import (
+        run_fastapi_server,
+        start_fastapi_server,
+    )
+
+    try:
+        result = start_fastapi_server(
+            port=args.port,
+            host=args.host,
+            db_path=args.db,
+            auto_port=args.auto_port,
+            reload=args.reload,
+        )
+
+        # Print server info
+        console.print("\n[bold cyan]HtmlGraph FastAPI Dashboard[/bold cyan]")
+        console.print("[bold green]✓[/bold green] Started observability dashboard")
+        console.print(f"URL: [bold blue]{result.handle.url}[/bold blue]")
+        console.print(f"Database: {result.config_used['db_path']}")
+
+        if result.warnings:
+            for warning in result.warnings:
+                console.print(f"[yellow]Warning: {warning}[/yellow]")
+
+        console.print("\n[cyan]Features:[/cyan]")
+        console.print("  • Real-time agent activity feed")
+        console.print("  • Orchestration chains visualization")
+        console.print("  • Feature tracker with Kanban view")
+        console.print("  • Session metrics & performance analytics")
+        console.print("  • WebSocket live event streaming")
+        console.print("\n[cyan]Press Ctrl+C to stop.[/cyan]\n")
+
+        # Run server
+        asyncio.run(run_fastapi_server(result.handle))
+
+    except KeyboardInterrupt:
+        console.print("\n[yellow]Shutting down...[/yellow]")
+    except Exception as e:
+        console.print(f"\n[red]Error:[/red] {e}")
+        import traceback
+
+        if getattr(args, "verbose", False):
+            traceback.print_exc()
+        sys.exit(1)
 
 
 def cmd_init(args: argparse.Namespace) -> None:
@@ -142,28 +254,26 @@ def cmd_init(args: argparse.Namespace) -> None:
 
     # Interactive setup wizard
     if args.interactive:
-        print("=== HtmlGraph Interactive Setup ===\n")
+        console.print("\n[bold cyan]=== HtmlGraph Interactive Setup ===[/bold cyan]\n")
 
         # Get project name
         default_name = Path(args.dir).resolve().name
-        project_name = input(f"Project name [{default_name}]: ").strip() or default_name
+        project_name = Prompt.ask("Project name", default=default_name)
 
         # Get agent name
-        agent_name = input("Your agent name [claude]: ").strip() or "claude"
+        agent_name = Prompt.ask("Your agent name", default="claude")
 
         # Ask about git hooks
-        install_hooks_response = (
-            input("Install git hooks for automatic tracking? [Y/n]: ").strip().lower()
+        args.install_hooks = Confirm.ask(
+            "Install git hooks for automatic tracking?", default=True
         )
-        args.install_hooks = install_hooks_response != "n"
 
         # Ask about documentation generation
-        gen_docs_response = (
-            input("Generate AGENTS.md, CLAUDE.md, GEMINI.md? [Y/n]: ").strip().lower()
+        generate_docs = Confirm.ask(
+            "Generate AGENTS.md, CLAUDE.md, GEMINI.md?", default=True
         )
-        generate_docs = gen_docs_response != "n"
 
-        print()
+        console.print()
     else:
         # Non-interactive defaults
         project_name = Path(args.dir).resolve().name
@@ -548,7 +658,7 @@ exit 0
             content = render_template(agents_template, replacements)
             if content:
                 agents_dest.write_text(content, encoding="utf-8")
-                print(f"✓ Generated: {agents_dest}")
+                console.print(f"[green]✓ Generated:[/green] {agents_dest}")
 
         # Generate CLAUDE.md
         claude_template = templates_dir / "CLAUDE.md.template"
@@ -557,7 +667,7 @@ exit 0
             content = render_template(claude_template, replacements)
             if content:
                 claude_dest.write_text(content, encoding="utf-8")
-                print(f"✓ Generated: {claude_dest}")
+                console.print(f"[green]✓ Generated:[/green] {claude_dest}")
 
         # Generate GEMINI.md
         gemini_template = templates_dir / "GEMINI.md.template"
@@ -566,15 +676,17 @@ exit 0
             content = render_template(gemini_template, replacements)
             if content:
                 gemini_dest.write_text(content, encoding="utf-8")
-                print(f"✓ Generated: {gemini_dest}")
+                console.print(f"[green]✓ Generated:[/green] {gemini_dest}")
 
     def install_hooks_step() -> None:
         if not args.install_hooks:
             return
         git_dir = Path(args.dir) / ".git"
         if not git_dir.exists():
-            print("\n⚠️  Warning: No .git directory found. Git hooks not installed.")
-            print("   Initialize git first: git init")
+            console.print(
+                "\n[yellow]⚠️  Warning: No .git directory found. Git hooks not installed.[/yellow]"
+            )
+            console.print("[dim]Initialize git first: git init[/dim]")
             return
 
         def install_hook(
@@ -607,11 +719,11 @@ exit 0
             git_hook_path = git_dir / "hooks" / hook_name
 
             if git_hook_path.exists():
-                print(f"\n⚠️  Existing {hook_name} hook found")
+                console.print(f"\n[yellow]⚠️  Existing {hook_name} hook found[/yellow]")
                 backup_path = git_hook_path.with_suffix(".existing")
                 if not backup_path.exists():
                     shutil.copy(git_hook_path, backup_path)
-                    print(f"   Backed up to: {backup_path}")
+                    console.print(f"[dim]Backed up to: {backup_path}[/dim]")
 
                 chain_content = f'''#!/bin/bash
 # Chained hook - runs existing hook then HtmlGraph hook
@@ -626,18 +738,18 @@ fi
 '''
                 git_hook_path.write_text(chain_content)
                 git_hook_path.chmod(0o755)
-                print(f"   Installed chained hook at: {git_hook_path}")
+                console.print(f"[dim]Installed chained hook at: {git_hook_path}[/dim]")
                 return
 
             try:
                 git_hook_path.symlink_to(hook_dest.resolve())
-                print("\n✓ Git hooks installed")
-                print(f"  {hook_name}: {git_hook_path} -> {hook_dest}")
+                console.print("\n[green]✓ Git hooks installed[/green]")
+                console.print(f"[dim]{hook_name}: {git_hook_path} -> {hook_dest}[/dim]")
             except OSError:
                 shutil.copy(hook_dest, git_hook_path)
                 git_hook_path.chmod(0o755)
-                print("\n✓ Git hooks installed")
-                print(f"  {hook_name}: {git_hook_path}")
+                console.print("\n[green]✓ Git hooks installed[/green]")
+                console.print(f"[dim]{hook_name}: {git_hook_path}[/dim]")
 
         install_hook("pre-commit", hook_files["pre-commit"], pre_commit)
         install_hook("post-commit", hook_files["post-commit"], post_commit)
@@ -645,7 +757,9 @@ fi
         install_hook("post-merge", hook_files["post-merge"], post_merge)
         install_hook("pre-push", hook_files["pre-push"], pre_push)
 
-        print("\nGit events will now be logged to HtmlGraph automatically.")
+        console.print(
+            "\n[cyan]Git events will now be logged to HtmlGraph automatically.[/cyan]"
+        )
 
     steps: list[tuple[str, Any]] = [
         ("Create .htmlgraph directories", create_graph_dirs),
@@ -678,14 +792,19 @@ fi
 
     run_steps(steps)
 
-    print(f"\nInitialized HtmlGraph in {graph_dir}")
-    print(f"Collections: {', '.join(HtmlGraphAPIHandler.COLLECTIONS)}")
-    print("\nStart server with: htmlgraph serve")
+    console.print()
+    console.print(f"[green]✓ Initialized HtmlGraph in[/green] {graph_dir}")
+    console.print(
+        f"[cyan]Collections:[/cyan] {', '.join(HtmlGraphAPIHandler.COLLECTIONS)}"
+    )
+    console.print()
+    console.print("[bold cyan]Start server with:[/bold cyan]")
+    console.print("  htmlgraph serve", style="dim")
     if not args.no_index:
-        print(
-            f"Analytics cache: {graph_dir / 'index.sqlite'} (rebuildable; typically gitignored)"
+        console.print(
+            f"[dim]Analytics cache: {graph_dir / 'index.sqlite'} (rebuildable; typically gitignored)[/dim]"
         )
-    print(f"Events: {events_dir}/ (append-only JSONL)")
+    console.print(f"[dim]Events: {events_dir}/ (append-only JSONL)[/dim]")
 
 
 def cmd_install_hooks(args: argparse.Namespace) -> None:
@@ -704,22 +823,24 @@ def cmd_install_hooks(args: argparse.Namespace) -> None:
     # Handle configuration changes
     if args.enable:
         if args.enable not in AVAILABLE_HOOKS:
-            print(f"Error: Unknown hook '{args.enable}'")
-            print(f"Available hooks: {', '.join(AVAILABLE_HOOKS)}")
+            console.print(f"[red]Error: Unknown hook '{args.enable}'[/red]")
+            console.print(f"[cyan]Available hooks:[/cyan] {', '.join(AVAILABLE_HOOKS)}")
             return
         config.enable_hook(args.enable)
         config.save()
-        print(f"✓ Enabled hook '{args.enable}' in configuration")
+        console.print(f"[green]✓ Enabled hook '{args.enable}' in configuration[/green]")
         return
 
     if args.disable:
         if args.disable not in AVAILABLE_HOOKS:
-            print(f"Error: Unknown hook '{args.disable}'")
-            print(f"Available hooks: {', '.join(AVAILABLE_HOOKS)}")
+            console.print(f"[red]Error: Unknown hook '{args.disable}'[/red]")
+            console.print(f"[cyan]Available hooks:[/cyan] {', '.join(AVAILABLE_HOOKS)}")
             return
         config.disable_hook(args.disable)
         config.save()
-        print(f"✓ Disabled hook '{args.disable}' in configuration")
+        console.print(
+            f"[green]✓ Disabled hook '{args.disable}' in configuration[/green]"
+        )
         return
 
     # Override symlink preference if --use-copy is set
@@ -732,84 +853,116 @@ def cmd_install_hooks(args: argparse.Namespace) -> None:
     # Validate environment
     is_valid, error_msg = installer.validate_environment()
     if not is_valid:
-        print(f"❌ {error_msg}")
+        console.print(f"[red]❌ {error_msg}[/red]")
         return
 
     # List hooks status
     if args.list:
-        print("\nGit Hooks Installation Status")
-        print("=" * 60)
+        console.print()
+        table = Table(title="Git Hooks Installation Status", border_style="cyan")
+        table.add_column("Hook", style="cyan", no_wrap=True)
+        table.add_column("Enabled", justify="center", style="green")
+        table.add_column("Installed", justify="center", style="green")
+        table.add_column("Type", style="dim")
+        table.add_column("Status", style="dim")
 
         status = installer.list_hooks()
         for hook_name, info in status.items():
             status_icon = "✓" if info["installed"] else "✗"
             enabled_icon = "🟢" if info["enabled"] else "🔴"
+            type_str = "Symlink" if info["is_symlink"] else "Copied"
+            our_hook = (
+                "✓"
+                if info.get("our_hook", False)
+                else "✗"
+                if info["is_symlink"]
+                else ""
+            )
+
+            status_str = (
+                f"{type_str} {our_hook}".strip() if type_str == "Symlink" else type_str
+            )
+
+            table.add_row(hook_name, enabled_icon, status_icon, type_str, status_str)
 
-            print(f"\n{enabled_icon} {hook_name} ({status_icon} installed)")
-            print(f"  Enabled in config: {info['enabled']}")
-            print(f"  Versioned (.htmlgraph/hooks/): {info['versioned']}")
-            print(f"  Installed (.git/hooks/): {info['installed']}")
-
-            if info["is_symlink"]:
-                our_hook = "✓" if info.get("our_hook", False) else "✗"
-                print(f"  Type: Symlink ({our_hook} ours)")
-                print(f"  Target: {info.get('symlink_target', 'unknown')}")
-            elif info["installed"]:
-                print("  Type: Copied file")
-
-        print("\n" + "=" * 60)
-        print(f"\nConfiguration: {config_path}")
-        print("Use 'htmlgraph install-hooks --enable <hook>' to enable")
-        print("Use 'htmlgraph install-hooks --disable <hook>' to disable")
+        console.print(table)
+        console.print()
+        console.print(f"[dim]Configuration: {config_path}[/dim]")
+        console.print(
+            "[dim]Use 'htmlgraph install-hooks --enable <hook>' to enable[/dim]"
+        )
+        console.print(
+            "[dim]Use 'htmlgraph install-hooks --disable <hook>' to disable[/dim]"
+        )
         return
 
     # Uninstall a hook
     if args.uninstall:
         if args.uninstall not in AVAILABLE_HOOKS:
-            print(f"Error: Unknown hook '{args.uninstall}'")
-            print(f"Available hooks: {', '.join(AVAILABLE_HOOKS)}")
+            console.print(f"[red]Error: Unknown hook '{args.uninstall}'[/red]")
+            console.print(f"[cyan]Available hooks:[/cyan] {', '.join(AVAILABLE_HOOKS)}")
             return
 
         success, message = installer.uninstall_hook(args.uninstall)
         if success:
-            print(f"✓ {message}")
+            console.print(f"[green]✓ {message}[/green]")
         else:
-            print(f"❌ {message}")
+            console.print(f"[red]❌ {message}[/red]")
         return
 
     # Install hooks
-    print("\n🔧 Installing Git hooks for HtmlGraph\n")
-    print(f"Project: {project_dir}")
-    print(f"Configuration: {config_path}")
+    console.print("\n[cyan]🔧 Installing Git hooks for HtmlGraph[/cyan]\n")
+    console.print(f"[dim]Project: {project_dir}[/dim]")
+    console.print(f"[dim]Configuration: {config_path}[/dim]")
 
     if args.dry_run:
-        print("\n[DRY RUN MODE - No changes will be made]\n")
+        console.print("\n[yellow][DRY RUN MODE - No changes will be made][/yellow]\n")
 
     results = installer.install_all_hooks(force=args.force, dry_run=args.dry_run)
 
-    # Display results
+    # Display results in a table
+    result_table = Table(title="Installation Results", border_style="cyan")
+    result_table.add_column("Hook", style="cyan", no_wrap=True)
+    result_table.add_column("Status", justify="center")
+    result_table.add_column("Message", style="dim")
+
     success_count = 0
     failure_count = 0
 
     for hook_name, (success, message) in results.items():
         if success:
             success_count += 1
-            print(f"✓ {message}")
+            result_table.add_row(hook_name, "[green]✓ Success[/green]", message)
         else:
             failure_count += 1
-            print(f"❌ {message}")
+            result_table.add_row(hook_name, "[red]❌ Failed[/red]", message)
 
-    print("\n" + "=" * 60)
-    print(f"Summary: {success_count} installed, {failure_count} failed")
+    console.print(result_table)
+    console.print()
+    console.print(
+        f"[cyan]Summary:[/cyan] {success_count} installed, {failure_count} failed"
+    )
 
     if not args.dry_run:
-        print(f"\nConfiguration saved to: {config_path}")
-        print("\nGit events will now be logged to HtmlGraph automatically.")
-        print("\nManagement commands:")
-        print("  htmlgraph install-hooks --list          # Show status")
-        print("  htmlgraph install-hooks --uninstall <hook>  # Remove hook")
-        print("  htmlgraph install-hooks --enable <hook>     # Enable hook")
-        print("  htmlgraph install-hooks --disable <hook>    # Disable hook")
+        console.print()
+        console.print(f"[green]Configuration saved to: {config_path}[/green]")
+        console.print(
+            "[cyan]Git events will now be logged to HtmlGraph automatically.[/cyan]"
+        )
+        console.print()
+        console.print("[bold cyan]Management commands:[/bold cyan]")
+        console.print(
+            "[dim]  htmlgraph install-hooks --list          # Show status[/dim]"
+        )
+        console.print(
+            "[dim]  htmlgraph install-hooks --uninstall <hook>  # Remove hook[/dim]"
+        )
+        console.print(
+            "[dim]  htmlgraph install-hooks --enable <hook>     # Enable hook[/dim]"
+        )
+        console.print(
+            "[dim]  htmlgraph install-hooks --disable <hook>    # Disable hook[/dim]"
+        )
 
 
 def cmd_status(args: argparse.Namespace) -> None:
@@ -819,8 +972,9 @@ def cmd_status(args: argparse.Namespace) -> None:
 
     from htmlgraph.sdk import SDK
 
-    # Use SDK to query all collections
-    sdk = SDK(directory=args.graph_dir)
+    # Use SDK to query all collections with status spinner
+    with console.status("[blue]Initializing SDK...", spinner="dots"):
+        sdk = SDK(directory=args.graph_dir)
 
     total = 0
     by_status: Counter[str] = Counter()
@@ -839,22 +993,34 @@ def cmd_status(args: argparse.Namespace) -> None:
         "agents",
     ]
 
-    for coll_name in collections:
-        coll = getattr(sdk, coll_name)
-        try:
-            nodes = coll.all()
-            count = len(nodes)
-            if count > 0:
-                by_collection[coll_name] = count
-                total += count
-
-                # Count by status
-                for node in nodes:
-                    status = getattr(node, "status", "unknown")
-                    by_status[status] += 1
-        except Exception:
-            # Collection might not exist yet
-            pass
+    # Use progress bar for scanning collections
+    with Progress(
+        SpinnerColumn(),
+        TextColumn("[progress.description]{task.description}"),
+        console=console,
+        transient=True,
+    ) as progress:
+        task = progress.add_task("Scanning collections...", total=len(collections))
+
+        for coll_name in collections:
+            progress.update(task, description=f"Scanning {coll_name}...")
+            coll = getattr(sdk, coll_name)
+            try:
+                nodes = coll.all()
+                count = len(nodes)
+                if count > 0:
+                    by_collection[coll_name] = count
+                    total += count
+
+                    # Count by status
+                    for node in nodes:
+                        status = getattr(node, "status", "unknown")
+                        by_status[status] += 1
+            except Exception:
+                # Collection might not exist yet
+                pass
+
+            progress.update(task, advance=1)
 
     # Output based on format flag
     if args.format == "json":
@@ -995,17 +1161,21 @@ def cmd_query(args: argparse.Namespace) -> None:
 
     graph_dir = Path(args.graph_dir)
     if not graph_dir.exists():
-        print(f"Error: {graph_dir} not found.", file=sys.stderr)
+        console.print(f"[red]Error: {graph_dir} not found.[/red]")
         sys.exit(1)
 
-    results = []
-    for collection_dir in graph_dir.iterdir():
-        if collection_dir.is_dir() and not collection_dir.name.startswith("."):
-            graph = HtmlGraph(collection_dir, auto_load=True)
-            for node in graph.query(args.selector):
-                data = node_to_dict(node)
-                data["_collection"] = collection_dir.name
-                results.append(data)
+    # Query with status spinner
+    with console.status(
+        f"[blue]Querying with selector '{args.selector}'...", spinner="dots"
+    ):
+        results = []
+        for collection_dir in graph_dir.iterdir():
+            if collection_dir.is_dir() and not collection_dir.name.startswith("."):
+                graph = HtmlGraph(collection_dir, auto_load=True)
+                for node in graph.query(args.selector):
+                    data = node_to_dict(node)
+                    data["_collection"] = collection_dir.name
+                    results.append(data)
 
     if args.format == "json":
         print(json.dumps(results, indent=2, default=str))
@@ -1027,10 +1197,31 @@ def cmd_session_start(args: argparse.Namespace) -> None:
     """Start a new session."""
     import json
 
+    from pydantic import ValidationError
+
+    from htmlgraph.pydantic_models import SessionStartInput
     from htmlgraph.sdk import SDK
+    from htmlgraph.validation import display_validation_error
 
-    sdk = SDK(directory=args.graph_dir, agent=args.agent)
-    session = sdk.start_session(session_id=args.id, title=args.title, agent=args.agent)
+    try:
+        input_data = SessionStartInput(
+            session_id=args.id,
+            title=args.title,
+            agent=args.agent,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
+        sys.exit(1)
+
+    with console.status(
+        "[blue]Initializing SDK and starting session...", spinner="dots"
+    ):
+        sdk = SDK(directory=args.graph_dir, agent=input_data.agent)
+        session = sdk.start_session(
+            session_id=input_data.session_id,
+            title=input_data.title,
+            agent=input_data.agent,
+        )
 
     if args.format == "json":
         from htmlgraph.converter import session_to_dict
@@ -1048,19 +1239,33 @@ def cmd_session_end(args: argparse.Namespace) -> None:
     """End a session."""
     import json
 
+    from pydantic import ValidationError
+
+    from htmlgraph.pydantic_models import SessionEndInput
     from htmlgraph.sdk import SDK
+    from htmlgraph.validation import display_validation_error
+
+    try:
+        input_data = SessionEndInput(
+            session_id=args.id,
+            notes=args.notes,
+            recommend=args.recommend,
+            blocker=args.blocker if args.blocker else None,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
+        sys.exit(1)
 
     sdk = SDK(directory=args.graph_dir)
-    blockers = args.blocker if args.blocker else None
     session = sdk.end_session(
-        args.id,
-        handoff_notes=args.notes,
-        recommended_next=args.recommend,
-        blockers=blockers,
+        input_data.session_id,
+        handoff_notes=input_data.notes,
+        recommended_next=input_data.recommend,
+        blockers=input_data.blocker,
     )
 
     if session is None:
-        print(f"Error: Session '{args.id}' not found.", file=sys.stderr)
+        console.print(f"[red]Error: Session '{input_data.session_id}' not found.[/red]")
         sys.exit(1)
 
     if args.format == "json":
@@ -1079,18 +1284,30 @@ def cmd_session_handoff(args: argparse.Namespace) -> None:
     """Set or show session handoff context."""
     import json
 
+    from pydantic import ValidationError
+
+    from htmlgraph.pydantic_models import SessionHandoffInput
     from htmlgraph.sdk import SDK
+    from htmlgraph.validation import display_validation_error
 
-    sdk = SDK(directory=args.graph_dir, agent=args.agent)
+    try:
+        input_data = SessionHandoffInput(
+            session_id=args.session_id,
+            notes=args.notes,
+            recommend=args.recommend,
+            blocker=args.blocker if args.blocker else None,
+            show=args.show,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
+        sys.exit(1)
 
-    if args.show:
-        # For showing, we might still need direct manager access or add more methods to SDK
-        # But for now, let's keep using SessionManager logic via SDK property if needed
-        # or implement show logic here using SDK collections
+    sdk = SDK(directory=args.graph_dir, agent=args.agent)
 
+    if input_data.show:
         # Use session_manager.get_session() to get Session objects (not Node)
-        if args.session_id:
-            session = sdk.session_manager.get_session(args.session_id)
+        if input_data.session_id:
+            session = sdk.session_manager.get_session(input_data.session_id)
         else:
             # Need "last ended session" - SDK doesn't expose this yet.
             # Fallback to session_manager logic exposed on SDK
@@ -1118,7 +1335,7 @@ def cmd_session_handoff(args: argparse.Namespace) -> None:
         return
 
     # Setting handoff
-    if not (args.notes or args.recommend or args.blocker):
+    if not (input_data.notes or input_data.recommend or input_data.blocker):
         print(
             "Error: Provide --notes, --recommend, or --blocker (or use --show).",
             file=sys.stderr,
@@ -1126,15 +1343,17 @@ def cmd_session_handoff(args: argparse.Namespace) -> None:
         sys.exit(1)
 
     handoff_result = sdk.set_session_handoff(
-        session_id=args.session_id,  # Optional, defaults to active
-        handoff_notes=args.notes,
-        recommended_next=args.recommend,
-        blockers=args.blocker if args.blocker else None,
+        session_id=input_data.session_id,  # Optional, defaults to active
+        handoff_notes=input_data.notes,
+        recommended_next=input_data.recommend,
+        blockers=input_data.blocker,
     )
 
     if handoff_result is None:
-        if args.session_id:
-            print(f"Error: Session '{args.session_id}' not found.", file=sys.stderr)
+        if input_data.session_id:
+            print(
+                f"Error: Session '{input_data.session_id}' not found.", file=sys.stderr
+            )
         else:
             print(
                 "Error: No active session found. Provide --session-id.",
@@ -1160,17 +1379,20 @@ def cmd_session_list(args: argparse.Namespace) -> None:
         return
 
     converter = SessionConverter(sessions_dir)
-    sessions = converter.load_all()
 
-    # Sort by started_at descending (handle mixed tz-aware/naive datetimes)
-    def sort_key(s: Any) -> Any:
-        ts = s.started_at
-        # Make naive datetimes comparable by assuming UTC
-        if ts.tzinfo is None:
-            return ts.replace(tzinfo=None)
-        return ts.replace(tzinfo=None)  # Compare as naive for sorting
+    # Load sessions with status spinner
+    with console.status("[blue]Loading sessions...", spinner="dots"):
+        sessions = converter.load_all()
+
+        # Sort by started_at descending (handle mixed tz-aware/naive datetimes)
+        def sort_key(s: Any) -> Any:
+            ts = s.started_at
+            # Make naive datetimes comparable by assuming UTC
+            if ts.tzinfo is None:
+                return ts.replace(tzinfo=None)
+            return ts.replace(tzinfo=None)  # Compare as naive for sorting
 
-    sessions.sort(key=sort_key, reverse=True)
+        sessions.sort(key=sort_key, reverse=True)
 
     if args.format == "json":
         from htmlgraph.converter import session_to_dict
@@ -1178,17 +1400,34 @@ def cmd_session_list(args: argparse.Namespace) -> None:
         print(json.dumps([session_to_dict(s) for s in sessions], indent=2))
     else:
         if not sessions:
-            print("No sessions found.")
+            console.print("[yellow]No sessions found.[/yellow]")
             return
 
-        print(f"{'ID':<30} {'Status':<10} {'Agent':<15} {'Events':<8} {'Started'}")
-        print("=" * 90)
+        # Create Rich table
+        table = Table(
+            title="Sessions",
+            show_header=True,
+            header_style="bold magenta",
+            box=box.ROUNDED,
+        )
+        table.add_column("ID", style="cyan", no_wrap=False, max_width=30)
+        table.add_column("Status", style="green", width=10)
+        table.add_column("Agent", style="blue", width=15)
+        table.add_column("Events", justify="right", style="yellow", width=8)
+        table.add_column("Started", style="white")
+
         for session in sessions:
             started = session.started_at.strftime("%Y-%m-%d %H:%M")
-            print(
-                f"{session.id:<30} {session.status:<10} {session.agent:<15} {session.event_count:<8} {started}"
+            table.add_row(
+                session.id,
+                session.status,
+                session.agent,
+                str(session.event_count),
+                started,
             )
 
+        console.print(table)
+
 
 def cmd_session_start_info(args: argparse.Namespace) -> None:
     """Get comprehensive session start information (optimized for AI agents)."""
@@ -1210,16 +1449,18 @@ def cmd_session_start_info(args: argparse.Namespace) -> None:
     else:
         # Human-readable format
         status: dict = info["status"]  # type: ignore
-        print("=" * 80)
-        print("SESSION START INFO")
-        print("=" * 80)
-
-        # Project status
-        print(f"\nProject: {status.get('project_name', 'HtmlGraph')}")
-        print(f"Total features: {status.get('total_features', 0)}")
-        print(f"In progress: {status.get('wip_count', 0)}")
+
+        # Project status panel
         by_status = status.get("by_status", {})
-        print(f"Completed: {by_status.get('done', 0)}")
+        project_info = (
+            f"Project: {status.get('project_name', 'HtmlGraph')}\n"
+            f"Total features: {status.get('total_features', 0)}\n"
+            f"In progress: {status.get('wip_count', 0)}\n"
+            f"Completed: {by_status.get('done', 0)}"
+        )
+        console.print(
+            Panel(project_info, title="SESSION START INFO", border_style="cyan")
+        )
 
         # Active work item (validation status)
         active_work = info.get("active_work")
@@ -1382,13 +1623,17 @@ def cmd_session_dedupe(args: argparse.Namespace) -> None:
     """Move low-signal session files out of the main sessions directory."""
     from htmlgraph import SDK
 
-    sdk = SDK(directory=args.graph_dir)
-    result = sdk.dedupe_sessions(
-        max_events=args.max_events,
-        move_dir_name=args.move_dir,
-        dry_run=args.dry_run,
-        stale_extra_active=not args.no_stale_active,
-    )
+    with console.status("[blue]Initializing SDK...", spinner="dots"):
+        sdk = SDK(directory=args.graph_dir)
+
+    operation = "Analyzing" if args.dry_run else "Deduplicating"
+    with console.status(f"[blue]{operation} sessions...", spinner="dots"):
+        result = sdk.dedupe_sessions(
+            max_events=args.max_events,
+            move_dir_name=args.move_dir,
+            dry_run=args.dry_run,
+            stale_extra_active=not args.no_stale_active,
+        )
 
     print(f"Scanned: {result['scanned']}")
     print(f"Moved:   {result['moved']}")
@@ -1424,15 +1669,14 @@ def cmd_session_link(args: argparse.Namespace) -> None:
     session_graph = HtmlGraph(sessions_dir)
     session = session_graph.get(args.session_id)
     if not session:
-        print(f"Error: Failed to load session '{args.session_id}'", file=sys.stderr)
+        console.print(f"[red]Error: Failed to load session '{args.session_id}'[/red]")
         sys.exit(1)
 
     # Load feature
     feature_file = feature_dir / f"{args.feature_id}.html"
     if not feature_file.exists():
-        print(
-            f"Error: Feature '{args.feature_id}' not found at {feature_file}",
-            file=sys.stderr,
+        console.print(
+            f"[red]Error: Feature '{args.feature_id}' not found at {feature_file}[/red]"
         )
         sys.exit(1)
 
@@ -1620,6 +1864,70 @@ def cmd_session_validate_attribution(args: argparse.Namespace) -> None:
                 )
 
 
+def cmd_session_debug(args: argparse.Namespace) -> None:
+    """Show full error traceback and debugging information for a session."""
+    from htmlgraph.session_manager import SessionManager
+
+    manager = SessionManager(args.graph_dir)
+
+    try:
+        session = manager.get_session(args.session_id)
+    except Exception:
+        session = None
+
+    if not session:
+        console.print(f"[red]✗ Session not found:[/red] {args.session_id}")
+        sys.exit(1)
+
+    # Display session header
+    console.print("\n[bold cyan]Session Debug Information[/bold cyan]")
+    console.print(f"[dim]Session ID:[/dim] {session.id}")
+    console.print(f"[dim]Agent:[/dim] {session.agent}")
+    console.print(f"[dim]Status:[/dim] {session.status}")
+    console.print(
+        f"[dim]Started:[/dim] {session.started_at.strftime('%Y-%m-%d %H:%M:%S')}"
+    )
+
+    # Check for errors
+    errors = session.error_log
+    if not errors:
+        console.print("\n[green]✓ No errors in session[/green]")
+        return
+
+    console.print(f"\n[bold yellow]Errors ({len(errors)})[/bold yellow]")
+
+    for i, error in enumerate(errors, 1):
+        # Error header
+        console.print(
+            f"\n[bold]Error {i}[/bold] [{error.timestamp.strftime('%H:%M:%S')}]"
+        )
+        console.print(f"[red]{error.error_type}[/red]: {error.message}")
+
+        # Tool information
+        if error.tool:
+            console.print(f"[dim]Tool:[/dim] {error.tool}")
+
+        # Context information
+        if error.context:
+            console.print(f"[dim]Context:[/dim] {error.context}")
+
+        # Full traceback
+        if error.traceback:
+            console.print("\n[dim]Traceback:[/dim]")
+            console.print(f"[dim]{error.traceback}[/dim]")
+
+        console.print("[dim]─" * 80 + "[/dim]")
+
+    # Summary
+    console.print("\n[bold cyan]Summary[/bold cyan]")
+    console.print(f"[dim]Total errors:[/dim] {len(errors)}")
+    error_types: dict[str, int] = {}
+    for error in errors:
+        error_types[error.error_type] = error_types.get(error.error_type, 0) + 1
+    for error_type, count in sorted(error_types.items()):
+        console.print(f"[dim]  - {error_type}:[/dim] {count}")
+
+
 # =========================================================================
 # Transcript Commands
 # =========================================================================
@@ -1974,27 +2282,33 @@ def cmd_transcript_patterns(args: argparse.Namespace) -> None:
             )
         )
     else:
-        print("Workflow Patterns Detected")
-        print("=" * 50)
-
         optimal = [p for p in patterns if p.category == "optimal"]
         anti = [p for p in patterns if p.category == "anti-pattern"]
         neutral = [p for p in patterns if p.category == "neutral"][:10]
 
+        content = ""
         if optimal:
-            print("\n✅ Optimal Patterns:")
+            content += "✅ Optimal Patterns:\n"
             for p in optimal:
-                print(f"   {' → '.join(p.sequence)} ({p.count}x)")
+                content += f"   {' → '.join(p.sequence)} ({p.count}x)\n"
 
         if anti:
-            print("\n⚠️ Anti-Patterns:")
+            content += "\n⚠️  Anti-Patterns:\n"
             for p in anti:
-                print(f"   {' → '.join(p.sequence)} ({p.count}x)")
+                content += f"   {' → '.join(p.sequence)} ({p.count}x)\n"
 
         if neutral:
-            print("\n📊 Common Patterns:")
+            content += "\n📊 Common Patterns:\n"
             for p in neutral:
-                print(f"   {' → '.join(p.sequence)} ({p.count}x)")
+                content += f"   {' → '.join(p.sequence)} ({p.count}x)\n"
+
+        console.print(
+            Panel(
+                content.strip(),
+                title="Workflow Patterns Detected",
+                border_style="green",
+            )
+        )
 
 
 def cmd_transcript_transitions(args: argparse.Namespace) -> None:
@@ -2009,11 +2323,6 @@ def cmd_transcript_transitions(args: argparse.Namespace) -> None:
     if args.format == "json":
         print(json.dumps(transitions, indent=2))
     else:
-        print("Tool Transition Matrix")
-        print("=" * 50)
-        print("(from_tool → to_tool: count)")
-        print()
-
         # Flatten and sort
         flat = []
         for from_tool, tos in transitions.items():
@@ -2022,9 +2331,23 @@ def cmd_transcript_transitions(args: argparse.Namespace) -> None:
 
         flat.sort(key=lambda x: -x[2])
 
+        # Create Rich table
+        table = Table(
+            title="Tool Transition Matrix",
+            show_header=True,
+            header_style="bold magenta",
+            box=box.ROUNDED,
+        )
+        table.add_column("From Tool", style="cyan", width=12)
+        table.add_column("To Tool", style="blue", width=12)
+        table.add_column("Count", justify="right", style="yellow", width=6)
+        table.add_column("Visual", style="green")
+
         for from_t, to_t, count in flat[:20]:
             bar = "█" * min(count, 20)
-            print(f"  {from_t:12} → {to_t:12} {count:4} {bar}")
+            table.add_row(from_t, to_t, str(count), bar)
+
+        console.print(table)
 
 
 def cmd_transcript_recommendations(args: argparse.Namespace) -> None:
@@ -2041,10 +2364,10 @@ def cmd_transcript_recommendations(args: argparse.Namespace) -> None:
     if args.format == "json":
         print(json.dumps({"recommendations": recommendations}, indent=2))
     else:
-        print("Workflow Recommendations")
-        print("=" * 50)
-        for rec in recommendations:
-            print(f"  {rec}")
+        content = "\n".join([f"  • {rec}" for rec in recommendations])
+        console.print(
+            Panel(content, title="Workflow Recommendations", border_style="yellow")
+        )
 
 
 def cmd_transcript_insights(args: argparse.Namespace) -> None:
@@ -2071,24 +2394,43 @@ def cmd_transcript_insights(args: argparse.Namespace) -> None:
             )
         )
     else:
-        print("📊 Transcript Insights")
-        print("=" * 50)
-        print(f"Sessions Analyzed: {insights.total_sessions}")
-        print(f"Total User Messages: {insights.total_user_messages}")
-        print(f"Total Tool Calls: {insights.total_tool_calls}")
-        print(f"Avg Session Health: {insights.avg_session_health:.0%}")
-        print()
+        # Summary panel
+        summary = (
+            f"Sessions Analyzed: {insights.total_sessions}\n"
+            f"Total User Messages: {insights.total_user_messages}\n"
+            f"Total Tool Calls: {insights.total_tool_calls}\n"
+            f"Avg Session Health: {insights.avg_session_health:.0%}"
+        )
+        console.print(
+            Panel(summary, title="📊 Transcript Insights", border_style="cyan")
+        )
 
+        # Top tools table
         if insights.tool_frequency:
-            print("🔧 Top Tools:")
+            table = Table(
+                title="🔧 Top Tools",
+                show_header=True,
+                header_style="bold magenta",
+                box=box.ROUNDED,
+            )
+            table.add_column("Tool", style="cyan", width=15)
+            table.add_column("Count", justify="right", style="yellow", width=6)
+            table.add_column("Visual", style="green")
+
             for tool, count in list(insights.tool_frequency.items())[:8]:
                 bar = "█" * min(count // 5, 15)
-                print(f"   {tool:15} {count:4} {bar}")
+                table.add_row(tool, str(count), bar)
 
-        print()
-        print("💡 Recommendations:")
-        for rec in insights.recommendations[:5]:
-            print(f"   {rec}")
+            console.print(table)
+
+        # Recommendations panel
+        if insights.recommendations:
+            rec_content = "\n".join(
+                [f"   {rec}" for rec in insights.recommendations[:5]]
+            )
+            console.print(
+                Panel(rec_content, title="💡 Recommendations", border_style="yellow")
+            )
 
 
 def cmd_transcript_export(args: argparse.Namespace) -> None:
@@ -2210,18 +2552,34 @@ def cmd_track(args: argparse.Namespace) -> None:
     """Track an activity in the current session."""
     import json
 
+    from pydantic import ValidationError
+
     from htmlgraph import SDK
+    from htmlgraph.pydantic_models import ActivityTrackInput
+    from htmlgraph.validation import display_validation_error
+
+    try:
+        input_data = ActivityTrackInput(
+            tool=args.tool,
+            summary=args.summary,
+            files=args.files,
+            session=args.session,
+            failed=args.failed,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
+        sys.exit(1)
 
     agent = os.environ.get("HTMLGRAPH_AGENT")
     sdk = SDK(directory=args.graph_dir, agent=agent)
 
     try:
         entry = sdk.track_activity(
-            tool=args.tool,
-            summary=args.summary,
-            file_paths=args.files,
-            success=not args.failed,
-            session_id=args.session,  # None if not specified, SDK will find active session
+            tool=input_data.tool,
+            summary=input_data.summary,
+            file_paths=input_data.files,
+            success=not input_data.failed,
+            session_id=input_data.session,  # None if not specified, SDK will find active session
         )
     except ValueError as e:
         print(f"Error: {e}", file=sys.stderr)
@@ -2664,160 +3022,121 @@ def cmd_agent_list(args: argparse.Namespace) -> None:
 
 def cmd_feature_create(args: argparse.Namespace) -> None:
     """Create a new feature."""
-    import json
-
-    from htmlgraph.sdk import SDK
+    from pydantic import ValidationError
 
-    # Use SDK for feature creation (which now handles logging)
-    sdk = SDK(directory=args.graph_dir, agent=args.agent)
+    from htmlgraph.cli_commands.feature import FeatureCreateCommand
+    from htmlgraph.pydantic_models import FeatureCreateInput
+    from htmlgraph.validation import display_validation_error
 
     try:
-        # Determine collection (features -> create builder, others -> manual create?)
-        # For now, only 'features' has a builder in SDK.features.create()
-        # But BaseCollection doesn't have create().
-
-        # If collection is 'features', use builder
-        if args.collection == "features":
-            builder = sdk.features.create(
-                title=args.title,
-                description=args.description or "",
-                priority=args.priority,
-            )
-            if args.steps:
-                builder.add_steps(args.steps)
-            node = builder.save()
-        else:
-            # Fallback to SessionManager directly for non-feature collections
-            # (or extend SDK to support create on all collections)
-            # For consistency with old CLI, we use SessionManager here if not features.
-            # But wait, SDK initializes SessionManager.
-
-            # Creating bugs/chores via SDK isn't fully fluent yet.
-            # Let's use the low-level SessionManager.create_feature logic for now via SDK's session_manager
-            # IF we want to strictly use SDK. But SDK.session_manager IS exposed now.
-            node = sdk.session_manager.create_feature(
-                title=args.title,
-                collection=args.collection,
-                description=args.description or "",
-                priority=args.priority,
-                steps=args.steps,
-                agent=args.agent,
-            )
-
-    except ValueError as e:
-        print(f"Error: {e}", file=sys.stderr)
+        input_data = FeatureCreateInput(
+            title=args.title,
+            description=args.description,
+            priority=args.priority,
+            steps=args.steps,
+            collection=args.collection,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
         sys.exit(1)
 
-    if args.format == "json":
-        from htmlgraph.converter import node_to_dict
+    # Convert steps count to list of step names (e.g., 3 -> ["Step 1", "Step 2", "Step 3"])
+    step_names = None
+    if input_data.steps:
+        step_names = [f"Step {i + 1}" for i in range(input_data.steps)]
 
-        print(json.dumps(node_to_dict(node), indent=2))
-    else:
-        print(f"Created: {node.id}")
-        print(f"  Title: {node.title}")
-        print(f"  Status: {node.status}")
-        print(f"  Path: {args.graph_dir}/{args.collection}/{node.id}.html")
+    command = FeatureCreateCommand(
+        collection=input_data.collection,
+        title=input_data.title,
+        description=input_data.description or "",
+        priority=input_data.priority,
+        steps=step_names,
+        track_id=args.track,
+    )
+    command.run(graph_dir=args.graph_dir, agent=args.agent, output_format=args.format)
 
 
 def cmd_feature_start(args: argparse.Namespace) -> None:
     """Start working on a feature."""
-    import json
+    from pydantic import ValidationError
 
-    from htmlgraph.sdk import SDK
-
-    sdk = SDK(directory=args.graph_dir, agent=args.agent)
-    collection = getattr(sdk, args.collection, None)
-
-    if not collection:
-        print(
-            f"Error: Collection '{args.collection}' not found in SDK.", file=sys.stderr
-        )
-        sys.exit(1)
+    from htmlgraph.cli_commands.feature import FeatureStartCommand
+    from htmlgraph.pydantic_models import FeatureStartInput
+    from htmlgraph.validation import display_validation_error
 
     try:
-        node = collection.start(args.id)
-    except ValueError as e:
-        print(f"Error: {e}", file=sys.stderr)
-        sys.exit(1)
-
-    if node is None:
-        print(
-            f"Error: Feature '{args.id}' not found in {args.collection}.",
-            file=sys.stderr,
+        input_data = FeatureStartInput(
+            feature_id=args.id,
+            collection=args.collection,
         )
+    except ValidationError as e:
+        display_validation_error(e)
         sys.exit(1)
 
-    if args.format == "json":
-        from htmlgraph.converter import node_to_dict
-
-        print(json.dumps(node_to_dict(node), indent=2))
-    else:
-        print(f"Started: {node.id}")
-        print(f"  Title: {node.title}")
-        print(f"  Status: {node.status}")
-
-        # Show WIP status
-        status = sdk.session_manager.get_status()
-        print(f"  WIP: {status['wip_count']}/{status['wip_limit']}")
+    command = FeatureStartCommand(
+        collection=input_data.collection,
+        feature_id=input_data.feature_id,
+    )
+    command.run(graph_dir=args.graph_dir, agent=args.agent, output_format=args.format)
 
 
 def cmd_feature_complete(args: argparse.Namespace) -> None:
     """Mark a feature as complete."""
-    import json
+    from pydantic import ValidationError
 
-    from htmlgraph.sdk import SDK
-
-    sdk = SDK(directory=args.graph_dir, agent=args.agent)
-    collection = getattr(sdk, args.collection, None)
-
-    if not collection:
-        print(
-            f"Error: Collection '{args.collection}' not found in SDK.", file=sys.stderr
-        )
-        sys.exit(1)
+    from htmlgraph.cli_commands.feature import FeatureCompleteCommand
+    from htmlgraph.pydantic_models import FeatureCompleteInput
+    from htmlgraph.validation import display_validation_error
 
     try:
-        node = collection.complete(args.id)
-    except ValueError as e:
-        print(f"Error: {e}", file=sys.stderr)
-        sys.exit(1)
-
-    if node is None:
-        print(
-            f"Error: Feature '{args.id}' not found in {args.collection}.",
-            file=sys.stderr,
+        input_data = FeatureCompleteInput(
+            feature_id=args.id,
+            collection=args.collection,
         )
+    except ValidationError as e:
+        display_validation_error(e)
         sys.exit(1)
 
-    if args.format == "json":
-        from htmlgraph.converter import node_to_dict
-
-        print(json.dumps(node_to_dict(node), indent=2))
-    else:
-        print(f"Completed: {node.id}")
-        print(f"  Title: {node.title}")
+    command = FeatureCompleteCommand(
+        collection=input_data.collection,
+        feature_id=input_data.feature_id,
+    )
+    command.run(graph_dir=args.graph_dir, agent=args.agent, output_format=args.format)
 
 
 def cmd_feature_primary(args: argparse.Namespace) -> None:
     """Set the primary feature for attribution."""
     import json
 
+    from pydantic import ValidationError
+
+    from htmlgraph.pydantic_models import FeaturePrimaryInput
     from htmlgraph.sdk import SDK
+    from htmlgraph.validation import display_validation_error
+
+    try:
+        input_data = FeaturePrimaryInput(
+            feature_id=args.id,
+            collection=args.collection,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
+        sys.exit(1)
 
     sdk = SDK(directory=args.graph_dir, agent=args.agent)
 
     # Only FeatureCollection has set_primary currently
-    if args.collection == "features":
-        node = sdk.features.set_primary(args.id)
+    if input_data.collection == "features":
+        node = sdk.features.set_primary(input_data.feature_id)
     else:
         # Fallback to direct session manager for other collections
         node = sdk.session_manager.set_primary_feature(
-            args.id, collection=args.collection, agent=args.agent
+            input_data.feature_id, collection=input_data.collection, agent=args.agent
         )
 
     if node is None:
         print(
-            f"Error: Feature '{args.id}' not found in {args.collection}.",
+            f"Error: Feature '{input_data.feature_id}' not found in {input_data.collection}.",
             file=sys.stderr,
         )
         sys.exit(1)
@@ -2835,26 +3154,40 @@ def cmd_feature_claim(args: argparse.Namespace) -> None:
     """Claim a feature."""
     import json
 
+    from pydantic import ValidationError
+
+    from htmlgraph.pydantic_models import FeatureClaimInput
     from htmlgraph.sdk import SDK
+    from htmlgraph.validation import display_validation_error
+
+    try:
+        input_data = FeatureClaimInput(
+            feature_id=args.id,
+            collection=args.collection,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
+        sys.exit(1)
 
     sdk = SDK(directory=args.graph_dir, agent=args.agent)
-    collection = getattr(sdk, args.collection, None)
+    collection = getattr(sdk, input_data.collection, None)
 
     if not collection:
         print(
-            f"Error: Collection '{args.collection}' not found in SDK.", file=sys.stderr
+            f"Error: Collection '{input_data.collection}' not found in SDK.",
+            file=sys.stderr,
         )
         sys.exit(1)
 
     try:
-        node = collection.claim(args.id)
+        node = collection.claim(input_data.feature_id)
     except ValueError as e:
         print(f"Error: {e}", file=sys.stderr)
         sys.exit(1)
 
     if node is None:
         print(
-            f"Error: Feature '{args.id}' not found in {args.collection}.",
+            f"Error: Feature '{input_data.feature_id}' not found in {input_data.collection}.",
             file=sys.stderr,
         )
         sys.exit(1)
@@ -2873,26 +3206,40 @@ def cmd_feature_release(args: argparse.Namespace) -> None:
     """Release a feature."""
     import json
 
+    from pydantic import ValidationError
+
+    from htmlgraph.pydantic_models import FeatureReleaseInput
     from htmlgraph.sdk import SDK
+    from htmlgraph.validation import display_validation_error
+
+    try:
+        input_data = FeatureReleaseInput(
+            feature_id=args.id,
+            collection=args.collection,
+        )
+    except ValidationError as e:
+        display_validation_error(e)
+        sys.exit(1)
 
     sdk = SDK(directory=args.graph_dir, agent=args.agent)
-    collection = getattr(sdk, args.collection, None)
+    collection = getattr(sdk, input_data.collection, None)
 
     if not collection:
         print(
-            f"Error: Collection '{args.collection}' not found in SDK.", file=sys.stderr
+            f"Error: Collection '{input_data.collection}' not found in SDK.",
+            file=sys.stderr,
         )
         sys.exit(1)
 
     try:
-        node = collection.release(args.id)
+        node = collection.release(input_data.feature_id)
     except ValueError as e:
         print(f"Error: {e}", file=sys.stderr)
         sys.exit(1)
 
     if node is None:
         print(
-            f"Error: Feature '{args.id}' not found in {args.collection}.",
+            f"Error: Feature '{input_data.feature_id}' not found in {input_data.collection}.",
             file=sys.stderr,
         )
         sys.exit(1)
@@ -3193,11 +3540,10 @@ def cmd_cigs_reset_violations(args: argparse.Namespace) -> None:
 
     # Confirm reset
     if not args.yes:
-        print(f"Current violations: {summary.total_violations}")
-        print(f"Total waste: {summary.total_waste_tokens} tokens")
-        response = input("\nReset violations for current session? [y/N]: ")
-        if response.lower() not in ("y", "yes"):
-            print("Reset cancelled")
+        console.print(f"Current violations: {summary.total_violations}")
+        console.print(f"Total waste: {summary.total_waste_tokens} tokens")
+        if not Confirm.ask("\nReset violations for current session?", default=False):
+            console.print("Reset cancelled")
             return
 
     # Clear violations file
@@ -3208,6 +3554,552 @@ def cmd_cigs_reset_violations(args: argparse.Namespace) -> None:
     print("Starting fresh for this session")
 
 
+def cmd_cigs_cost_dashboard(args: argparse.Namespace) -> None:
+    """Generate cost attribution dashboard from HtmlGraph events."""
+    import webbrowser
+    from pathlib import Path
+
+    from htmlgraph.cigs.cost import CostCalculator
+    from htmlgraph.operations.events import query_events
+
+    graph_dir = Path(args.graph_dir or ".htmlgraph")
+
+    # Parse options
+    save = getattr(args, "save", False)
+    open_browser = getattr(args, "open", False)
+    output_json = getattr(args, "json", False)
+    output_path = getattr(args, "output", None)
+
+    # Display progress
+    with console.status("[blue]Analyzing HtmlGraph events...[/blue]", spinner="dots"):
+        try:
+            # Query all events
+            result = query_events(graph_dir=graph_dir, limit=None)
+            events = result.events if hasattr(result, "events") else []
+
+            if not events:
+                console.print(
+                    "[yellow]No events found. Run some work to generate analytics![/yellow]"
+                )
+                return
+
+            # Calculate costs from events
+            cost_calc = CostCalculator()
+            cost_summary = _analyze_event_costs(events, cost_calc)
+
+        except Exception as e:
+            console.print(f"[red]Error analyzing events: {e}[/red]")
+            return
+
+    # Generate output
+    if output_json:
+        _output_cost_json(cost_summary, output_path)
+    else:
+        html_content = _generate_cost_dashboard_html(cost_summary)
+
+        if save or output_path:
+            output_file = (
+                Path(output_path) if output_path else graph_dir / "cost-dashboard.html"
+            )
+            output_file.write_text(html_content)
+            console.print(f"[green]✓ Dashboard saved to: {output_file}[/green]")
+
+            if open_browser:
+                webbrowser.open(f"file://{output_file.absolute()}")
+                console.print("[blue]Opening dashboard in browser...[/blue]")
+        else:
+            # Display summary to console
+            _display_cost_summary(cost_summary)
+
+    # Print recommendations
+    _print_cost_recommendations(cost_summary)
+
+
+def _analyze_event_costs(events: list[dict], cost_calc: object) -> dict:
+    """Analyze events and calculate cost attribution."""
+    cost_summary: dict[str, Any] = {
+        "total_cost_tokens": 0,
+        "total_events": len(events),
+        "tool_costs": {},
+        "session_costs": {},
+        "delegation_count": 0,
+        "direct_execution_count": 0,
+        "cost_by_category": {},
+    }
+
+    for event in events:
+        try:
+            tool = event.get("tool", "unknown")
+            session_id = event.get("session_id", "unknown")
+            cost = (
+                event.get("predicted_tokens", 0)
+                or event.get("actual_tokens", 0)
+                or 2000
+            )
+
+            # Track by tool
+            if tool not in cost_summary["tool_costs"]:
+                cost_summary["tool_costs"][tool] = {"count": 0, "total_tokens": 0}
+            cost_summary["tool_costs"][tool]["count"] += 1
+            cost_summary["tool_costs"][tool]["total_tokens"] += cost
+
+            # Track by session
+            if session_id not in cost_summary["session_costs"]:
+                cost_summary["session_costs"][session_id] = {
+                    "count": 0,
+                    "total_tokens": 0,
+                }
+            cost_summary["session_costs"][session_id]["count"] += 1
+            cost_summary["session_costs"][session_id]["total_tokens"] += cost
+
+            # Track delegation vs direct
+            if tool in ["Task", "spawn_gemini", "spawn_codex", "spawn_copilot"]:
+                cost_summary["delegation_count"] += 1
+                category = "delegation"
+            else:
+                cost_summary["direct_execution_count"] += 1
+                category = "direct"
+
+            if category not in cost_summary["cost_by_category"]:
+                cost_summary["cost_by_category"][category] = {
+                    "count": 0,
+                    "total_tokens": 0,
+                }
+            cost_summary["cost_by_category"][category]["count"] += 1
+            cost_summary["cost_by_category"][category]["total_tokens"] += cost
+
+            cost_summary["total_cost_tokens"] += cost
+        except Exception:
+            continue
+
+    return cost_summary
+
+
+def _generate_cost_dashboard_html(cost_summary: dict) -> str:
+    """Generate HTML dashboard for cost attribution."""
+    from datetime import datetime
+
+    # Calculate metrics
+    total_cost = cost_summary["total_cost_tokens"]
+    total_events = cost_summary["total_events"]
+    avg_cost = total_cost / total_events if total_events > 0 else 0
+    delegation_pct = (
+        cost_summary["delegation_count"] / total_events * 100 if total_events > 0 else 0
+    )
+
+    # Estimate cost in dollars (assuming $0.001 per 1K tokens for simplicity)
+    cost_usd = total_cost / 1_000_000 * 5  # Rough estimate
+
+    # Sort tools by cost
+    sorted_tools = sorted(
+        cost_summary["tool_costs"].items(),
+        key=lambda x: x[1]["total_tokens"],
+        reverse=True,
+    )
+
+    # Sort sessions by cost
+    sorted_sessions = sorted(
+        cost_summary["session_costs"].items(),
+        key=lambda x: x[1]["total_tokens"],
+        reverse=True,
+    )
+
+    # Build tool cost rows
+    tool_rows = "".join(
+        f"""
+    <tr>
+        <td class="cell">{tool}</td>
+        <td class="cell number">{data["count"]}</td>
+        <td class="cell number">{data["total_tokens"]:,}</td>
+        <td class="cell number">{data["total_tokens"] / total_cost * 100:.1f}%</td>
+    </tr>
+    """
+        for tool, data in sorted_tools[:20]
+    )
+
+    # Build session cost rows
+    session_rows = "".join(
+        f"""
+    <tr>
+        <td class="cell">{session[:12]}...</td>
+        <td class="cell number">{data["count"]}</td>
+        <td class="cell number">{data["total_tokens"]:,}</td>
+        <td class="cell number">{data["total_tokens"] / total_cost * 100:.1f}%</td>
+    </tr>
+    """
+        for session, data in sorted_sessions[:20]
+    )
+
+    html = f"""<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>HtmlGraph Cost Dashboard</title>
+    <style>
+        * {{
+            margin: 0;
+            padding: 0;
+            box-sizing: border-box;
+        }}
+
+        body {{
+            font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
+            background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+            min-height: 100vh;
+            padding: 40px 20px;
+        }}
+
+        .container {{
+            max-width: 1400px;
+            margin: 0 auto;
+        }}
+
+        header {{
+            background: white;
+            border-radius: 12px;
+            padding: 30px;
+            margin-bottom: 30px;
+            box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
+        }}
+
+        h1 {{
+            color: #667eea;
+            margin-bottom: 10px;
+            font-size: 28px;
+        }}
+
+        .timestamp {{
+            color: #999;
+            font-size: 12px;
+        }}
+
+        .metrics {{
+            display: grid;
+            grid-template-columns: repeat(auto-fit, minmax(250px, 1fr));
+            gap: 20px;
+            margin-top: 20px;
+        }}
+
+        .metric {{
+            background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+            color: white;
+            padding: 20px;
+            border-radius: 8px;
+            box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
+        }}
+
+        .metric-label {{
+            font-size: 12px;
+            opacity: 0.9;
+            margin-bottom: 8px;
+            text-transform: uppercase;
+            letter-spacing: 1px;
+        }}
+
+        .metric-value {{
+            font-size: 32px;
+            font-weight: bold;
+        }}
+
+        .metric-unit {{
+            font-size: 14px;
+            opacity: 0.8;
+            margin-left: 8px;
+        }}
+
+        .metric.warning {{
+            background: linear-gradient(135deg, #f093fb 0%, #f5576c 100%);
+        }}
+
+        .metric.success {{
+            background: linear-gradient(135deg, #4facfe 0%, #00f2fe 100%);
+        }}
+
+        section {{
+            background: white;
+            border-radius: 12px;
+            padding: 30px;
+            margin-bottom: 30px;
+            box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
+        }}
+
+        h2 {{
+            color: #333;
+            margin-bottom: 20px;
+            font-size: 20px;
+            border-bottom: 2px solid #667eea;
+            padding-bottom: 10px;
+        }}
+
+        table {{
+            width: 100%;
+            border-collapse: collapse;
+        }}
+
+        th {{
+            background: #f5f5f5;
+            padding: 12px;
+            text-align: left;
+            font-weight: 600;
+            color: #333;
+            border-bottom: 2px solid #ddd;
+        }}
+
+        td {{
+            padding: 12px;
+            border-bottom: 1px solid #eee;
+        }}
+
+        td.cell {{
+            color: #333;
+        }}
+
+        td.number {{
+            text-align: right;
+            font-family: 'Monaco', 'Courier New', monospace;
+            color: #667eea;
+            font-weight: 500;
+        }}
+
+        tr:hover {{
+            background: #f9f9f9;
+        }}
+
+        .insights {{
+            display: grid;
+            grid-template-columns: repeat(auto-fit, minmax(300px, 1fr));
+            gap: 20px;
+            margin-top: 20px;
+        }}
+
+        .insight {{
+            background: #f0f4ff;
+            border-left: 4px solid #667eea;
+            padding: 16px;
+            border-radius: 4px;
+        }}
+
+        .insight-title {{
+            font-weight: 600;
+            color: #333;
+            margin-bottom: 8px;
+        }}
+
+        .insight-text {{
+            color: #666;
+            font-size: 14px;
+            line-height: 1.6;
+        }}
+
+        .footer {{
+            text-align: center;
+            color: #999;
+            font-size: 12px;
+            margin-top: 40px;
+        }}
+    </style>
+</head>
+<body>
+    <div class="container">
+        <header>
+            <h1>💰 HtmlGraph Cost Dashboard</h1>
+            <p class="timestamp">Generated: {datetime.now().strftime("%Y-%m-%d %H:%M:%S")}</p>
+            <div class="metrics">
+                <div class="metric">
+                    <div class="metric-label">Total Cost</div>
+                    <div class="metric-value">{total_cost:,}<span class="metric-unit">tokens</span></div>
+                </div>
+                <div class="metric success">
+                    <div class="metric-label">Estimated USD</div>
+                    <div class="metric-value">${cost_usd:.2f}</div>
+                </div>
+                <div class="metric">
+                    <div class="metric-label">Average Cost</div>
+                    <div class="metric-value">{avg_cost:,.0f}<span class="metric-unit">tokens</span></div>
+                </div>
+                <div class="metric success">
+                    <div class="metric-label">Delegation Rate</div>
+                    <div class="metric-value">{delegation_pct:.1f}%</div>
+                </div>
+            </div>
+        </header>
+
+        <section>
+            <h2>📊 Cost by Tool</h2>
+            <table>
+                <thead>
+                    <tr>
+                        <th>Tool</th>
+                        <th>Count</th>
+                        <th>Total Tokens</th>
+                        <th>% of Total</th>
+                    </tr>
+                </thead>
+                <tbody>
+                    {tool_rows}
+                </tbody>
+            </table>
+        </section>
+
+        <section>
+            <h2>🔄 Cost by Session</h2>
+            <table>
+                <thead>
+                    <tr>
+                        <th>Session ID</th>
+                        <th>Count</th>
+                        <th>Total Tokens</th>
+                        <th>% of Total</th>
+                    </tr>
+                </thead>
+                <tbody>
+                    {session_rows}
+                </tbody>
+            </table>
+        </section>
+
+        <section>
+            <h2>💡 Insights & Recommendations</h2>
+            <div class="insights">
+                <div class="insight">
+                    <div class="insight-title">✓ Delegation Usage</div>
+                    <div class="insight-text">
+                        You're delegating {delegation_pct:.1f}% of operations.
+                        {"Continue delegation for cost efficiency!" if delegation_pct > 50 else "Consider delegating more operations to reduce costs."}
+                    </div>
+                </div>
+                <div class="insight">
+                    <div class="insight-title">🎯 Top Cost Driver</div>
+                    <div class="insight-text">
+                        {sorted_tools[0][0] if sorted_tools else "N/A"} accounts for {sorted_tools[0][1]["total_tokens"] / total_cost * 100:.1f}% of total cost.
+                        Review if this tool usage is optimal.
+                    </div>
+                </div>
+                <div class="insight">
+                    <div class="insight-title">📈 Parallelization Opportunity</div>
+                    <div class="insight-text">
+                        Parallel Task() calls can reduce overall execution time by ~40%.
+                        Look for independent operations that can run simultaneously.
+                    </div>
+                </div>
+            </div>
+        </section>
+
+        <div class="footer">
+            <p>HtmlGraph Cost Attribution Dashboard | Real-time cost tracking and optimization</p>
+        </div>
+    </div>
+</body>
+</html>"""
+
+    return html
+
+
+def _output_cost_json(cost_summary: dict, output_path: str | None) -> None:
+    """Output cost data as JSON."""
+    import json
+
+    output_file = Path(output_path) if output_path else Path("cost-summary.json")
+    output_file.write_text(json.dumps(cost_summary, indent=2))
+    console.print(f"[green]✓ JSON output saved to: {output_file}[/green]")
+
+
+def _display_cost_summary(cost_summary: dict) -> None:
+    """Display cost summary in console."""
+    console.print("\n[bold cyan]Cost Dashboard Summary[/bold cyan]\n")
+
+    # Create summary table
+    summary_table = Table(
+        show_header=True, header_style="bold magenta", box=box.ROUNDED
+    )
+    summary_table.add_column("Metric", style="cyan")
+    summary_table.add_column("Value", style="green")
+
+    total_tokens = cost_summary["total_cost_tokens"]
+    total_events = cost_summary["total_events"]
+    avg_tokens = total_tokens / total_events if total_events > 0 else 0
+    delegation_pct = (
+        cost_summary["delegation_count"] / total_events * 100 if total_events > 0 else 0
+    )
+    cost_usd = total_tokens / 1_000_000 * 5
+
+    summary_table.add_row("Total Events", str(total_events))
+    summary_table.add_row("Total Cost", f"{total_tokens:,} tokens")
+    summary_table.add_row("Average Cost", f"{avg_tokens:,.0f} tokens/event")
+    summary_table.add_row("Estimated USD", f"${cost_usd:.2f}")
+    summary_table.add_row("Delegation Count", str(cost_summary["delegation_count"]))
+    summary_table.add_row("Delegation Rate", f"{delegation_pct:.1f}%")
+    summary_table.add_row(
+        "Direct Executions", str(cost_summary["direct_execution_count"])
+    )
+
+    console.print(summary_table)
+
+    # Top tools
+    if cost_summary["tool_costs"]:
+        console.print("\n[bold cyan]Top Cost Drivers (by Tool)[/bold cyan]\n")
+        tools_table = Table(
+            show_header=True, header_style="bold magenta", box=box.ROUNDED
+        )
+        tools_table.add_column("Tool", style="cyan")
+        tools_table.add_column("Count", justify="right", style="green")
+        tools_table.add_column("Tokens", justify="right", style="yellow")
+        tools_table.add_column("% Total", justify="right", style="magenta")
+
+        for tool, data in sorted(
+            cost_summary["tool_costs"].items(),
+            key=lambda x: x[1]["total_tokens"],
+            reverse=True,
+        )[:10]:
+            pct = data["total_tokens"] / total_tokens * 100
+            tools_table.add_row(
+                tool, str(data["count"]), f"{data['total_tokens']:,}", f"{pct:.1f}%"
+            )
+
+        console.print(tools_table)
+
+
+def _print_cost_recommendations(cost_summary: dict) -> None:
+    """Print recommendations for cost optimization."""
+    console.print("\n[bold cyan]Recommendations[/bold cyan]\n")
+
+    total_events = cost_summary["total_events"]
+    delegation_pct = (
+        cost_summary["delegation_count"] / total_events * 100 if total_events > 0 else 0
+    )
+
+    recommendations = []
+
+    if delegation_pct < 50:
+        recommendations.append(
+            "[yellow]→ Increase delegation usage[/yellow] - Consider using Task() and spawn_* for more operations"
+        )
+
+    if cost_summary["tool_costs"]:
+        top_tool = max(
+            cost_summary["tool_costs"].items(), key=lambda x: x[1]["total_tokens"]
+        )
+        if top_tool[1]["total_tokens"] / cost_summary["total_cost_tokens"] > 0.4:
+            recommendations.append(
+                f"[yellow]→ Review {top_tool[0]} usage[/yellow] - It accounts for {top_tool[1]['total_tokens'] / cost_summary['total_cost_tokens'] * 100:.1f}% of total cost"
+            )
+
+    if total_events > 100:
+        recommendations.append(
+            "[green]✓ Good event volume[/green] - Sufficient data for optimization analysis"
+        )
+
+    recommendations.append(
+        "[blue]💡 Tip: Use parallel Task() calls to reduce execution time by ~40%[/blue]"
+    )
+
+    for rec in recommendations:
+        console.print(f"  {rec}")
+
+    console.print()
+
+
 def cmd_publish(args: argparse.Namespace) -> None:
     """Build and publish the package to PyPI (Interoperable)."""
     import shutil
@@ -3228,14 +4120,14 @@ def cmd_publish(args: argparse.Namespace) -> None:
         shutil.rmtree(dist_dir)
 
     # 2. Build
-    print("Building package with uv...")
+    console.print("Building package with uv...", style="blue")
     try:
         subprocess.run(["uv", "build"], check=True)
     except subprocess.CalledProcessError:
-        print("Error: Build failed.", file=sys.stderr)
+        console.print("[red]Error: Build failed.[/red]")
         sys.exit(1)
     except FileNotFoundError:
-        print("Error: 'uv' command not found.", file=sys.stderr)
+        console.print("[red]Error: 'uv' command not found.[/red]")
         sys.exit(1)
 
     # 3. Publish
@@ -3316,39 +4208,61 @@ def cmd_feature_list(args: argparse.Namespace) -> None:
     else:
         if not nodes:
             if not args.quiet:
-                print(
-                    f"No features found with status '{args.status}'."
+                console.print(
+                    f"[yellow]No features found with status '{args.status}'.[/yellow]"
                     if args.status
-                    else "No features found."
+                    else "[yellow]No features found.[/yellow]"
                 )
             return
 
-        # Header (skip if quiet)
+        # Create Rich table (skip if quiet)
         if not args.quiet:
-            print(f"{'ID':<25} {'Status':<12} {'Priority':<10} {'Title'}")
-            print("=" * 80)
+            table = Table(
+                title="Features",
+                show_header=True,
+                header_style="bold magenta",
+                box=box.ROUNDED,
+            )
+            table.add_column("ID", style="cyan", no_wrap=True, width=25)
+            table.add_column("Status", style="green", width=12)
+            table.add_column("Priority", style="yellow", width=10)
+            table.add_column("Title", style="white")
 
-        # List features
-        for node in nodes:
-            title = node.title[:35] + "..." if len(node.title) > 38 else node.title
-            print(f"{node.id:<25} {node.status:<12} {node.priority:<10} {title}")
+            # List features
+            for node in nodes:
+                title = node.title[:35] + "..." if len(node.title) > 38 else node.title
+                table.add_row(node.id, node.status, node.priority or "-", title)
 
-        # Verbose output
+            console.print(table)
+        else:
+            # Quiet mode - simple output without table
+            for node in nodes:
+                print(f"{node.id}\t{node.status}\t{node.priority}\t{node.title}")
+
+        # Verbose output with Rich.Panel
         if args.verbose >= 1:
-            print("\n--- Verbose Details ---")
-            print(f"Total features: {len(nodes)}")
-            print(f"Graph directory: {args.graph_dir}")
+            details = f"Total features: {len(nodes)}\nGraph directory: {args.graph_dir}"
             if args.status:
-                print(f"Filtered by status: {args.status}")
+                details += f"\nFiltered by status: {args.status}"
+
+            console.print(Panel(details, title="Verbose Details", border_style="cyan"))
 
         if args.verbose >= 2:
-            print("\nFeature breakdown by status:")
             from collections import Counter
 
             status_counts = Counter(n.status for n in sdk.features.all())
-            for status, count in sorted(status_counts.items()):
-                marker = "→" if status == args.status else " "
-                print(f"  {marker} {status}: {count}")
+            breakdown = "\n".join(
+                [
+                    f"  {'→' if status == args.status else ' '} {status}: {count}"
+                    for status, count in sorted(status_counts.items())
+                ]
+            )
+
+            console.print(
+                Panel(
+                    breakdown, title="Feature Breakdown by Status", border_style="blue"
+                )
+            )
 
 
 # =============================================================================
@@ -3377,7 +4291,7 @@ def cmd_feature_step_complete(args: argparse.Namespace) -> None:
     step_indices = sorted(set(step_indices))
 
     if not step_indices:
-        print("Error: No step indices provided", file=sys.stderr)
+        console.print("[red]Error: No step indices provided[/red]")
         sys.exit(1)
 
     # Make API requests for each step
@@ -3463,14 +4377,13 @@ def cmd_feature_delete(args: argparse.Namespace) -> None:
 
     # Confirmation prompt (unless --yes flag)
     if not args.yes:
-        print(f"Delete {args.collection.rstrip('s')} '{args.id}'?")
-        print(f"  Title: {feature.title}")
-        print(f"  Status: {feature.status}")
-        print("\nThis cannot be undone. Continue? [y/N] ", end="")
+        console.print(f"Delete {args.collection.rstrip('s')} '{args.id}'?")
+        console.print(f"  Title: {feature.title}")
+        console.print(f"  Status: {feature.status}")
+        console.print("\n[bold red]This cannot be undone.[/bold red]")
 
-        response = input().strip().lower()
-        if response not in ("y", "yes"):
-            print("Cancelled")
+        if not Confirm.ask("Continue?", default=False):
+            console.print("Cancelled")
             sys.exit(0)
 
     # Delete
@@ -3545,12 +4458,23 @@ def cmd_track_list(args: argparse.Namespace) -> None:
         print(json.dumps({"tracks": track_ids}, indent=2))
     else:
         if not track_ids:
-            print("No tracks found.")
-            print("\nCreate a track with: htmlgraph track new 'Track Title'")
+            console.print("[yellow]No tracks found.[/yellow]")
+            console.print(
+                "\n[dim]Create a track with: htmlgraph track new 'Track Title'[/dim]"
+            )
             return
 
-        print(f"Tracks in {args.graph_dir}/tracks/:")
-        print("=" * 60)
+        # Create Rich table
+        table = Table(
+            title=f"Tracks in {args.graph_dir}/tracks/",
+            show_header=True,
+            header_style="bold magenta",
+            box=box.ROUNDED,
+        )
+        table.add_column("Track ID", style="cyan", no_wrap=True)
+        table.add_column("Components", style="green")
+        table.add_column("Format", style="blue")
+
         for track_id in track_ids:
             # Check for both consolidated (single file) and directory-based formats
             track_file = Path(args.graph_dir) / "tracks" / f"{track_id}.html"
@@ -3564,12 +4488,12 @@ def cmd_track_list(args: argparse.Namespace) -> None:
                     or 'data-section="requirements"' in content
                 )
                 has_plan = 'data-section="plan"' in content
-                format_indicator = " (consolidated)"
+                format_type = "consolidated"
             else:
                 # Directory format
                 has_spec = (track_dir / "spec.html").exists()
                 has_plan = (track_dir / "plan.html").exists()
-                format_indicator = ""
+                format_type = "directory"
 
             components = []
             if has_spec:
@@ -3577,8 +4501,11 @@ def cmd_track_list(args: argparse.Namespace) -> None:
             if has_plan:
                 components.append("plan")
 
-            components_str = f" [{', '.join(components)}]" if components else " [empty]"
-            print(f"  {track_id}{components_str}{format_indicator}")
+            components_str = ", ".join(components) if components else "empty"
+
+            table.add_row(track_id, components_str, format_type)
+
+        console.print(table)
 
 
 def cmd_track_spec(args: argparse.Namespace) -> None:
@@ -4221,6 +5148,32 @@ For more help: https://github.com/Shakes-tzd/htmlgraph
         help="Automatically find an available port if default is occupied",
     )
 
+    # serve-api (FastAPI-based dashboard with real-time observability)
+    serve_api_parser = subparsers.add_parser(
+        "serve-api",
+        help="Start the FastAPI-based observability dashboard (Phase 3)",
+    )
+    serve_api_parser.add_argument(
+        "--port", "-p", type=int, default=8000, help="Port (default: 8000)"
+    )
+    serve_api_parser.add_argument(
+        "--host", default="127.0.0.1", help="Host to bind to (default: 127.0.0.1)"
+    )
+    serve_api_parser.add_argument(
+        "--db", default=None, help="Path to SQLite database file"
+    )
+    serve_api_parser.add_argument(
+        "--auto-port",
+        action="store_true",
+        help="Automatically find an available port if default is occupied",
+    )
+    serve_api_parser.add_argument(
+        "--reload",
+        action="store_true",
+        help="Enable auto-reload on file changes (development mode)",
+    )
+    serve_api_parser.set_defaults(func=cmd_serve_api)
+
     # init
     init_parser = subparsers.add_parser("init", help="Initialize .htmlgraph directory")
     init_parser.add_argument(
@@ -4486,6 +5439,17 @@ For more help: https://github.com/Shakes-tzd/htmlgraph
     session_validate.add_argument(
         "--format", "-f", choices=["text", "json"], default="text", help="Output format"
     )
+    session_validate.set_defaults(func=cmd_session_validate_attribution)
+
+    # session debug
+    session_debug = session_subparsers.add_parser(
+        "debug", help="Show error tracebacks and debug information for a session"
+    )
+    session_debug.add_argument("session_id", help="Session ID to debug")
+    session_debug.add_argument(
+        "--graph-dir", "-g", default=".htmlgraph", help="Graph directory"
+    )
+    session_debug.set_defaults(func=cmd_session_debug)
 
     # activity (legacy: was "track")
     activity_parser = subparsers.add_parser(
@@ -4824,6 +5788,11 @@ For more help: https://github.com/Shakes-tzd/htmlgraph
         help="Priority",
     )
     feature_create.add_argument("--steps", nargs="*", help="Implementation steps")
+    feature_create.add_argument(
+        "--track",
+        "-t",
+        help="Track ID to link feature to (required for features collection)",
+    )
     feature_create.add_argument(
         "--agent",
         default=os.environ.get("HTMLGRAPH_AGENT") or "cli",
@@ -5565,6 +6534,27 @@ For more help: https://github.com/Shakes-tzd/htmlgraph
         "--graph-dir", "-g", default=".htmlgraph", help="Graph directory"
     )
 
+    # cigs cost dashboard
+    cigs_cost_dashboard = cigs_subparsers.add_parser(
+        "cost-dashboard", help="Generate cost attribution dashboard"
+    )
+    cigs_cost_dashboard.add_argument(
+        "--graph-dir", "-g", default=".htmlgraph", help="Graph directory"
+    )
+    cigs_cost_dashboard.add_argument(
+        "--save",
+        "-s",
+        action="store_true",
+        help="Save to .htmlgraph/cost-dashboard.html",
+    )
+    cigs_cost_dashboard.add_argument(
+        "--open", "-o", action="store_true", help="Open in browser after generation"
+    )
+    cigs_cost_dashboard.add_argument(
+        "--json", action="store_true", help="Output JSON instead of HTML"
+    )
+    cigs_cost_dashboard.add_argument("--output", type=str, help="Custom output path")
+
     # install-gemini-extension
     subparsers.add_parser(
         "install-gemini-extension",
@@ -5598,6 +6588,8 @@ For more help: https://github.com/Shakes-tzd/htmlgraph
 
     if args.command == "serve":
         cmd_serve(args)
+    elif args.command == "serve-api":
+        cmd_serve_api(args)
     elif args.command == "init":
         cmd_init(args)
     elif args.command == "install-hooks":
@@ -5840,6 +6832,8 @@ For more help: https://github.com/Shakes-tzd/htmlgraph
             cmd_cigs_patterns(args)
         elif args.cigs_command == "reset-violations":
             cmd_cigs_reset_violations(args)
+        elif args.cigs_command == "cost-dashboard":
+            cmd_cigs_cost_dashboard(args)
         else:
             cigs_parser.print_help()
             sys.exit(1)
@@ -6016,14 +7010,17 @@ def cmd_archive_create(args: argparse.Namespace) -> None:
         print(f"Error: Directory not found: {htmlgraph_dir}", file=sys.stderr)
         sys.exit(1)
 
-    manager = ArchiveManager(htmlgraph_dir)
+    with console.status("[blue]Initializing archive manager...", spinner="dots"):
+        manager = ArchiveManager(htmlgraph_dir)
 
-    # Run archive operation
-    result = manager.archive_entities(
-        older_than_days=args.older_than,
-        period=args.period,
-        dry_run=args.dry_run,
-    )
+    # Run archive operation with status spinner
+    operation = "Previewing" if args.dry_run else "Archiving"
+    with console.status(f"[blue]{operation} entities...", spinner="dots"):
+        result = manager.archive_entities(
+            older_than_days=args.older_than,
+            period=args.period,
+            dry_run=args.dry_run,
+        )
 
     if result["dry_run"]:
         print("\n🔍 DRY RUN - Preview (no changes made)\n")
@@ -6055,26 +7052,55 @@ def cmd_archive_search(args: argparse.Namespace) -> None:
         print(f"Error: Directory not found: {htmlgraph_dir}", file=sys.stderr)
         sys.exit(1)
 
-    manager = ArchiveManager(htmlgraph_dir)
+    with console.status("[blue]Initializing archive manager...", spinner="dots"):
+        manager = ArchiveManager(htmlgraph_dir)
 
-    # Search archives
-    results = manager.search(args.query, limit=args.limit)
+    # Search archives with status spinner
+    with console.status(
+        f"[blue]Searching archives for '{args.query}'...", spinner="dots"
+    ):
+        results = manager.search(args.query, limit=args.limit)
 
     if args.format == "json":
         print(json.dumps({"query": args.query, "results": results}, indent=2))
     else:
-        print(f"\n🔍 Search results for: '{args.query}'\n")
-        print(f"Found {len(results)} result(s):\n")
-
-        for i, result in enumerate(results, 1):
-            print(f"{i}. {result['entity_id']} ({result['entity_type']})")
-            print(f"   Archive: {result['archive_file']}")
-            print(f"   Status: {result['status']}")
-            print(f"   Title: {result['title_snippet']}")
-            if result["description_snippet"]:
-                print(f"   Description: {result['description_snippet']}")
-            print(f"   Relevance: {result['rank']:.2f}")
-            print()
+        console.print(f"\n🔍 Search results for: '{args.query}'\n")
+        console.print(f"Found {len(results)} result(s):\n")
+
+        # Use progress bar for large result sets
+        if len(results) > 10:
+            with Progress(
+                SpinnerColumn(),
+                TextColumn("[progress.description]{task.description}"),
+                console=console,
+                transient=True,
+            ) as progress:
+                task = progress.add_task("Displaying results...", total=len(results))
+                for i, result in enumerate(results, 1):
+                    console.print(
+                        f"{i}. {result['entity_id']} ({result['entity_type']})"
+                    )
+                    console.print(f"   Archive: {result['archive_file']}")
+                    console.print(f"   Status: {result['status']}")
+                    console.print(f"   Title: {result['title_snippet']}")
+                    if result["description_snippet"]:
+                        console.print(
+                            f"   Description: {result['description_snippet']}"
+                        )
+                    console.print(f"   Relevance: {result['rank']:.2f}")
+                    console.print()
+                    progress.update(task, advance=1)
+        else:
+            # No progress bar for small result sets
+            for i, result in enumerate(results, 1):
+                console.print(f"{i}. {result['entity_id']} ({result['entity_type']})")
+                console.print(f"   Archive: {result['archive_file']}")
+                console.print(f"   Status: {result['status']}")
+                console.print(f"   Title: {result['title_snippet']}")
+                if result["description_snippet"]:
+                    console.print(f"   Description: {result['description_snippet']}")
+                console.print(f"   Relevance: {result['rank']:.2f}")
+                console.print()
 
     manager.close()