concurrency

name	concurrency
description	Design concurrent and parallel systems correctly. Invoke for: "concurrency design", "thread safety", "parallel processing", "worker pool", "queue", "producer consumer", "concurrent writes", "data consistency with concurrent access".
argument-hint	concurrent system to design or review
allowed-tools	Read, Write, Edit, Glob

Skill: Concurrency — Parallel & Concurrent Systems

Category: Optimization/Research

Role

Design correct, efficient concurrent systems — thread pools, queues, locks, and lock-free patterns.

When to invoke

• Concurrent access to shared state
• Worker pool design
• Queue-based processing
• "make this parallel"

Instructions

1. Identify: what's shared state? What must be atomic?
2. Choose synchronization: mutex, semaphore, atomic, lock-free
3. Design: producer-consumer, worker pool, event-driven
4. Avoid: lock contention, deadlocks, starvation
5. Use: asyncio.gather for I/O-bound, ProcessPoolExecutor for CPU-bound
6. Test: run concurrent test cases to expose race conditions

Output format

# Worker pool pattern
async def process_batch(items: list) -> list:
    semaphore = asyncio.Semaphore(10)  # max 10 concurrent
    async def process_one(item):
        async with semaphore:
            return await expensive_operation(item)
    return await asyncio.gather(*[process_one(i) for i in items])

Example

/concurrency design worker pool for processing 1000 LLM requests with rate limiting