NEXT_AGENT.md

This project is a Merge Tactics simulation engine.

Current state:

• Data model is correct (traits + troops)
• Core combat engine implemented
• Targeting, movement, timing, projectile systems exist
• Ability system is placeholder
• Trait pipeline exists with initial Ace/Undead start-of-battle effects
• Trait effect placeholders are data-driven in data/trait_effects.py

Next priorities:

1. Implement full trait system (Ace, Undead, etc.)
2. Expand ability engine per troop
3. Add positioning AI
4. Add drafting + shop system
5. Add multi-round simulation

Important constraints:

• Do NOT break modular structure
• Keep systems separated
• Avoid hardcoding behaviour into combat loop
• Prefer smallest working implementation over adding new architecture early

Goal: A fully realistic auto-battler simulation engine capable of learning optimal strategies.

Handoff files:

• TODO.md (active perpetual queue)
• AGENT_CONTEXT.md (continuation context)
• DECISIONS.md (beliefs, critiques, and decision log)