persona_name = "Archi"

role = "Architect"

background = "Design"

specialties = ["S1", "S2"]

skill_names = ["file_skill"]

action_instructions = "WRITE_FILE: test.py\n```\ncontent\n```"

memory_section = "Memory"

system_prompt = f"""=== IDENTITY ===

You are {persona_name}, an expert AI agent in the Swarm V2 system.

Role: {role}

Background: {background}

Specialties: {', '.join(specialties)}

=== ACTIONS ===

When you need to take action (create files, search, etc.), you MUST emit these EXACT tags:

{action_instructions}

=== RULES ===

1. Rule 1

2. Rule 2

{memory_section}"""

print(f"Original System Prompt:\n{system_prompt}\n")

parts = system_prompt.split("===")

print(f"Parts Count: {len(parts)}")

for i, p in enumerate(parts):

print(f"Part {i}: '{p}'")

if len(parts) >= 5:

final_system = "=== IDENTITY ===" + parts[2] + "=== ACTIONS ===" + parts[4]

print(f"\nFinal System Prompt (Length {len(final_system)}):\n{final_system}")

else:

print("\nNOT ENOUGH PARTS")