refactor(orchestration): B3 merge enhanced_orchestration into orchestration (#10666)

[mrveiss]

Thinking Path

1. Inventory: Mapped all 8 modules in enhanced_orchestration/ against existing orchestration/ modules. Identified 2 symbol collisions (WorkflowPlan, ExecutionStrategy) and 1 naming conflict (StrategyPlanner vs WorkflowPlanner).
2. Symbol resolution:
   • WorkflowPlan: promoted the enhanced_orchestration subclass (adds structured_criteria) as a drop-in replacement in orchestration/types.py. IS-A _SharedWorkflowPlan so backward-compatible.
   • ExecutionStrategy: canonical type lives in autobot_shared.workflow; not duplicated — re-exported from orchestration/__init__ for convenience.
   • StrategyPlanner vs WorkflowPlanner: genuinely different classes; both kept with distinct names.
3. Wire-in: Moved all 8 modules into orchestration/ with updated imports. Verified orchestration/__init__.py exports all new symbols.
4. Delete: Removed enhanced_orchestration/ and tests/enhanced_orchestration/ only after verifying zero functional import statements remain.
5. Startup smoke: Fixed orchestration/workflow_planning.py importing ExecutionStrategy from .types (not there — must come from autobot_shared.workflow). Fixed test_workflow_integration.py same issue.

What Changed

New modules in orchestration/:

• agent_router.py — TaskAgentScorer / AgentRouter alias (GH discovery(orchestration): dual AgentRouter classes (522 LOC vs 68 LOC) with same name, different scope #6819)
• blocked_plan_resumer.py — Redis pub-sub skill-promotion subscriber (ADR-006 Phase 3 / feat(orchestration/planner): async gap-fill resume path for blocked plans (#7268 Phase 3) #7431)
• collaboration_coordinator.py — Redis pub-sub multi-agent collaboration (discovery(workflow_runner): collaboration layer (_ensure_redis/_coordinate_collaboration/_broadcast_to_agents) should be a separate CollaborationCoordinator collaborator #6393)
• execution_strategies/ — ExecutionStrategyHandler dispatcher + 5 strategy implementations (sequential/parallel/pipeline/collaborative/adaptive)
• subagent_dispatcher.py — Autonomous subagent spawning with LLM reflection pass
• workflow_planning.py — StrategyPlanner (multi-agent; distinct from WorkflowPlanner)
• workflow_runner.py — WorkflowRunner multi-agent execution engine (refactor(orchestrator): three overlapping workflow execution paths in single Orchestrator class #5058)

Updated existing orchestration/ files:

• types.py: +FALLBACK_TIERS, +WorkflowPlan subclass (structured_criteria), +WorkflowDependencies, +AgentTask alias
• __init__.py: re-exports all new symbols + ExecutionStrategy re-export
• performance_tracker.py: fix TYPE_CHECKING import to orchestration.types

Updated orchestrator.py: all enhanced_orchestration.* imports replaced with orchestration.* and autobot_shared.workflow.

Deleted:

• autobot-backend/enhanced_orchestration/ (25 files)
• autobot-backend/tests/enhanced_orchestration/ (2 files)

Tests migrated to tests/orchestration/ (10 files; all enhanced_orchestration.* imports updated).

autobot-infrastructure/shared/scripts/phase_validation_system.py: updated directory reference.

Verification

# Zero functional enhanced_orchestration imports remain
git grep -w "enhanced_orchestration" -- "*.py" | grep -E "^[^#]+:(from|import) enhanced_orchestration"
# → (no output)

# Startup import smoke: 339 passed, 0 failed
python3 -m pytest autobot-backend/tests/test_startup_imports.py -q
# → 339 passed, 82 warnings

# flake8 clean on all changed paths
python3 -m flake8 --config=.flake8 autobot-backend/orchestration/ autobot-backend/orchestrator.py autobot-backend/tests/orchestration/
# → exit 0

Model Used

Claude Sonnet 4.6

Part of #10666

[github-actions]

✅ SSOT Configuration Compliance: Passing

🎉 No hardcoded values detected that have SSOT config equivalents!