Summary

Implement the ThroughputAdaptiveStrategy for the pluggable ceremony scheduling system.

Ceremonies fire when the team's throughput rate changes significantly. Not "every N tasks" but "when pace anomaly detected." Like an automated burndown chart monitor.

Design Reference

See Ceremony Scheduling design page -- "Throughput-Adaptive" strategy section.

Key Behaviors

• should_fire_ceremony: checks if rolling throughput rate has dropped below or spiked above configured thresholds (e.g., standup when velocity drops 30% from rolling average)
• should_transition_sprint: uses task completion threshold (like task-driven) -- no anomaly-based transition
• on_sprint_activated: initializes throughput rate tracking window
• on_task_completed: updates rolling rate measurement (tasks per unit time over a configurable window)
• on_task_blocked: may indicate throughput stall -- factor into rate calculation
• Default velocity calculator: TaskDrivenVelocityCalculator with rate-of-change as secondary metric

Strategy Config

strategy_config:
  velocity_drop_threshold_pct: 30     # fire standup when velocity drops 30%
  velocity_spike_threshold_pct: 50    # fire review when velocity spikes 50%
  measurement_window_tasks: 10        # rolling window size for rate calculation

Implementation

• New: src/synthorg/engine/workflow/strategies/throughput_adaptive.py
• New: tests/unit/engine/workflow/strategies/test_throughput_adaptive.py
• Implement CeremonySchedulingStrategy protocol from feat: sprint ceremony runtime scheduler with task-driven cadence support #961
• This is a stateful strategy (uses lifecycle hooks to track rate)

Dependencies

• feat: sprint ceremony runtime scheduler with task-driven cadence support #961 (protocol interfaces + reference implementation)