Add LangChain DeepAgents example and documentation

Example application demonstrating the LangChain backend with sandbox
templates, skills definitions, and a kind cluster test script.

Includes README for the langchain-agent-sandbox package with usage
examples for direct sandbox, template-based, and warm pool modes.

Signed-off-by: Johann-Peter Hartmann <johann-peter.hartmann@mayflower.de>

Author: johannhartmann

clients/python/langchain-agent-sandbox/README.md

@@ -15,4 +15,5 @@ This directory contains examples of how to use the Agent Sandbox. Each subdirect
 - [**policy**](./policy): Examples of using different policies with sandboxes.
 - [**python-runtime-sandbox**](./python-runtime-sandbox): An example of a Python runtime sandbox.
 - [**sandbox-ksa**](./sandbox-ksa): Examples of a sandbox with a service account, namespace, and a basic sandbox.
-- [**vscode-sandbox**](./vscode-sandbox): An example of running VSCode in a sandbox.
+- [**vscode-sandbox**](./vscode-sandbox): An example of running VSCode in a sandbox.
+- [**langchain-deepagents**](./langchain-deepagents): LangChain DeepAgents with agent-sandbox backend for secure code execution.

examples/README.md

@@ -0,0 +1,41 @@
+---
+name: data-analysis
+description: "Analyze data files and generate insights. Trigger on phrases like 'analyze this data', 'show statistics', 'summarize the CSV'."
+version: 1.0.0
+tags: [data, analysis, statistics]
+---
+# Data Analysis
+
+## Overview
+This skill helps analyze data files (CSV, JSON, text) and generate statistical summaries.
+
+## Prerequisites
+The sandbox should have Python with pandas available, or use basic Python for simple analysis.
+
+## Workflow
+1. Read the data file to understand its structure
+2. Write an analysis script
+3. Execute and capture results
+4. Present findings in a clear format
+
+## Analysis Templates
+
+### CSV Summary
+```python
+import csv
+with open('/app/data.csv') as f:
+    reader = csv.DictReader(f)
+    rows = list(reader)
+    print(f"Rows: {len(rows)}")
+    print(f"Columns: {reader.fieldnames}")
+```
+
+### JSON Exploration
+```python
+import json
+with open('/app/data.json') as f:
+    data = json.load(f)
+    print(f"Type: {type(data)}")
+    if isinstance(data, list):
+        print(f"Items: {len(data)}")
+```

examples/langchain-deepagents/.deepagents/skills/data-analysis/SKILL.md

@@ -0,0 +1,51 @@
+---
+name: file-organizer
+description: "Organize files by type, date, or custom criteria. Trigger on phrases like 'organize files', 'sort by type', 'clean up directory'."
+version: 1.0.0
+tags: [files, organization, cleanup]
+---
+# File Organizer
+
+## Overview
+This skill helps organize files in the sandbox by moving them to appropriate directories based on type, name patterns, or custom criteria.
+
+## Workflow
+1. List files in the target directory with `ls_info`
+2. Categorize files based on extension or pattern
+3. Create destination directories
+4. Move files using shell commands
+
+## Organization Strategies
+
+### By Extension
+```python
+import os
+import shutil
+
+extensions = {
+    '.py': 'python',
+    '.js': 'javascript',
+    '.md': 'docs',
+    '.json': 'data',
+    '.csv': 'data',
+}
+
+for filename in os.listdir('/app'):
+    ext = os.path.splitext(filename)[1]
+    if ext in extensions:
+        dest = f'/app/{extensions[ext]}'
+        os.makedirs(dest, exist_ok=True)
+        shutil.move(f'/app/{filename}', f'{dest}/{filename}')
+```
+
+### By Date (modification time)
+```python
+import os
+from datetime import datetime
+
+for filename in os.listdir('/app'):
+    path = f'/app/{filename}'
+    mtime = os.path.getmtime(path)
+    date_str = datetime.fromtimestamp(mtime).strftime('%Y-%m')
+    # organize by year-month
+```

examples/langchain-deepagents/.deepagents/skills/file-organizer/SKILL.md

@@ -0,0 +1,39 @@
+---
+name: python-script
+description: "Generate and run Python scripts. Trigger on phrases like 'write a script', 'create a Python program', 'run Python code'."
+version: 1.0.0
+tags: [python, scripting, automation]
+---
+# Python Script Generation
+
+## Overview
+This skill helps you create, debug, and run Python scripts in the sandbox environment.
+
+## Workflow
+1. Create the script file using `write`
+2. Execute with `execute("python /app/script.py")`
+3. If errors occur, read the file, fix issues, and re-run
+
+## Best Practices
+- Always include error handling with try/except
+- Use descriptive variable names
+- Add comments for complex logic
+- Print results for verification
+
+## Common Patterns
+
+### File Processing
+```python
+import os
+for filename in os.listdir('/app/data'):
+    with open(f'/app/data/{filename}') as f:
+        # process file
+```
+
+### Data Transformation
+```python
+import json
+data = json.loads(input_string)
+# transform
+output = json.dumps(result, indent=2)
+```