feat: add activation handler system for object interactions

[daveey]

TL;DR

Added a new data-driven activation handler system that allows configuring object interactions through declarative configuration rather than hardcoded behavior.

What changed?

• Implemented a new activation handler system that allows objects to define how they respond when agents interact with them
• Added configuration structures for filters (VibeFilter, ResourceFilter) and mutations (ResourceDelta, ResourceTransfer, Alignment, Freeze, Attack)
• Integrated activation handlers with the Move action system to trigger when agents move onto objects
• Updated GridObject to support activation handlers and added helper methods to set and retrieve them
• Added comprehensive tests for the new activation handler system
• Updated the cogsguard recipe to use the new activation handlers for align/scramble functionality

How to test?

1. Run the new test file test_activation_handlers.py to verify all handler types work correctly
2. Try the updated cogsguard recipe which now uses activation handlers instead of dedicated align/scramble actions
3. Create custom activation handlers in Python config and verify they trigger when agents interact with objects

Why make this change?

This change makes the environment more flexible by allowing object interactions to be defined through configuration rather than hardcoded C++ behavior. This enables:

1. Creating new interaction types without modifying C++ code
2. Composing complex behaviors from simple building blocks
3. Defining context-sensitive interactions based on agent state (vibe, resources)
4. Reducing code duplication by centralizing interaction logic in a single system

The activation handler system will eventually replace the legacy vibe-specific action handlers, making the codebase more maintainable and easier to extend.

[daveey]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• feat: add station sprites to mettascope #4603 : 2 dependent PRs (#4590 , #4604 )
• feat: add gear stations that clear inventory and provide specific gear #4589
• feat: add activation handler system for object interactions #4588 👈 (View in Graphite)
• Add window state persistence and AOE visualization to MettaScope #4566
• feat: add AOE overlays and make Assembler implement Alignable interface #4587
• Add align action and update related action handlers #4565
• Add observer_agent_id parameter to obs_features for agent-specific observations #4539
• Add Area of Effect (AOE) resource effects to grid objects #4538
• Add CommonsChest for shared inventory access across agents #4537
• Add Commons system for shared inventories between grid objects #4536
• Rename cog_arena.py to cogsguard.py and remove unused imports #4477
• feat: enforce inventory limits in cpp #4602
• Enhance attack system with armor, weapons, and configurable loot #4313 : 1 other dependent PR (#4321 )
• Make Transfer action optional and add vibe-triggered attack #4315
• Implement bidirectional resource transfer system with vibe-based triggers #4314
• feat: support metta://policy/<short_name_or_class_path> URIs for built-in policies #4561
• feat: allow negative inventory_regen_amounts for decay #4564
• Add clarification about when to run install commands #4563
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.