# oxwhirl/smac

SMAC: The StarCraft Multi-Agent Challenge

Repository: https://github.com/oxwhirl/smac
Canonical: https://ross.abutalabs.com/products/smac
Language: Python
License: MIT
License Family: permissive
Topics: starcraft-ii, reinforcement-learning, multiagent-systems, benchmark, machine-learning
Last push: 2024-02-18T19:49:08+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2794, "days_push": 927, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1366, forks 242 (observed 2026-08-28T04:04:31.248721+00:00)

## What it is
SMAC is a benchmark environment for cooperative multi-agent reinforcement learning built on StarCraft II via Blizzard's SC2 ML API and DeepMind's PySC2. It provides decentralised micromanagement scenarios where each game unit is controlled by an individual RL agent.

## Use cases
- benchmark cooperative multi-agent reinforcement learning algorithms
- train decentralized RL agents in StarCraft II micromanagement scenarios
- evaluate algorithms like QMIX and COMA
- research environment for MARL papers
- compare MARL methods on standardized scenarios

## When to choose
- you need a standard benchmark for cooperative multi-agent RL
- you want decentralized control scenarios with discrete actions and observations
- you are reproducing MARL research like QMIX or COMA

## When to avoid
- you need single-agent RL environments
- you cannot install StarCraft II or the SC2 ML API
- you want the newest scenario generation - use SMACv2 instead

## Facets
- artifact type: dataset
- maturity: maintenance
- function: reinforcement-learning, simulation, benchmarking, machine-learning
- domain: reinforcement-learning, artificial-intelligence
- platform: windows, python
- tags: starcraft-ii, multi-agent-reinforcement-learning, marl, benchmark, pysc2, decentralized-micromanagement, qmix, coma, game-development, research, linux, macos

## Member repositories
- oxwhirl/smac (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:31.248721+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:41:13.791895+00:00, confidence not recorded.
  - readme: https://github.com/oxwhirl/smac (fetched 2026-08-28T04:04:31.248721+00:00, sha cd040459e34c)
- Data as of 2026-08-30T08:39:29.467469+00:00.