Python Execution Fallback

When to Use

Use this skill when execute_code_sandbox returns 'unknown error' or fails unexpectedly, even on simple Python code like print statements.

Solution

Switch to using run_shell with Python heredoc syntax instead of execute_code_sandbox.

Basic Syntax

python3 << 'EOF'
# Your Python code here
print("Hello World")
result = 2 + 2
print(f"Result: {result}")
EOF

How to Call

Instead of:

execute_code_sandbox(code="print('Hello')")

Use:

run_shell(command="python3 << 'EOF'\nprint('Hello')\nEOF")

For multi-line code, build the command string:

code = """
import json
data = {"status": "ok", "count": 42}
print(json.dumps(data))
"""
run_shell(command=f"python3 << 'EOF'\n{code}\nEOF")

Why This Works

• run_shell executes commands directly in the shell environment, bypassing sandbox issues
• Heredoc syntax (<< 'EOF') allows clean multi-line Python code without complex escaping
• Single quotes around EOF prevent unwanted shell variable expansion
• More reliable for simple to moderate Python execution tasks

Guidelines

1. Always use single quotes around EOF ('EOF') to prevent variable expansion
2. Test with a simple print statement first to verify Python is available
3. For longer scripts, consider writing to a .py file first, then executing:

   run_shell(command="python3 /path/to/script.py")

4. Capture stdout from run_shell to get Python output
5. Handle errors by checking run_shell's stderr output

Example: Complete Workflow

# Step 1: Try execute_code_sandbox
try:
    result = execute_code_sandbox(code="print(2+2)")
except Exception as e:
    # Step 2: Fallback to run_shell with heredoc
    result = run_shell(command="python3 << 'EOF'\nprint(2+2)\nEOF")
    # Parse stdout to get the output
    output = result.stdout.strip()

Troubleshooting