/reason

Grok Swarm — Reason

Multi-perspective collaborative reasoning using Grok 4.20's multi-agent swarm.

Usage

/grok-swarm:reason <question-or-task>

Examples

/grok-swarm:reason Compare microservices vs monolith for this project
/grok-swarm:reason What's the best database schema for a multi-tenant SaaS?
/grok-swarm:reason Should we use GraphQL or REST for this API?
/grok-swarm:reason Evaluate the trade-offs of our current auth approach

Step 1 — Check API key

python3 -c "
import json, os
from pathlib import Path
config = Path.home() / '.config' / 'grok-swarm' / 'config.json'
env_key = os.environ.get('OPENROUTER_API_KEY') or os.environ.get('XAI_API_KEY')
if env_key:
    print('key:env')
elif config.exists():
    data = json.loads(config.read_text())
    print('key:file' if data.get('api_key') else 'nokey')
else:
    print('nokey')
"

If nokey, direct the user to /grok-swarm:setup or /grok-swarm:set-key.

Step 2 — Run reasoning

PLUGIN_ROOT="$(find ~/.claude/plugins -name 'grok_bridge.py' -exec dirname {} \; 2>/dev/null | head -1 | xargs -I{} dirname {})"
[ -z "$PLUGIN_ROOT" ] && PLUGIN_ROOT="$(find /usr /usr/local ~/.local -name 'grok_bridge.py' -exec dirname {} \; 2>/dev/null | head -1 | xargs -I{} dirname {})"
python3 "$PLUGIN_ROOT/src/bridge/cli.py" reason "$ARGUMENTS"

Step 3 — Present results

Present the multi-agent consensus and highlight where agents disagreed — divergent perspectives are often the most valuable part of reason mode.