Configuration

Configuration

All configuration is done through environment variables. Copy the example files and edit as needed:

cp .env.example .env                    # Core agent
cp .env.swarm.example .env.swarm        # Swarm system

Core Agent

Variable	Default	Description
NEBULUS_BASE_URL	http://localhost:5000/v1	LLM server URL
NEBULUS_API_KEY	admin	API key for LLM server
NEBULUS_MODEL	Meta-Llama-3.1-8B-Instruct-exl2-8_0	Model name
NEBULUS_TIMEOUT	300	Request timeout in seconds
NEBULUS_STREAMING	true	Enable streaming responses
COMPOSE_PROJECT_NAME	nebulus-atom	Docker Compose project name

Slack Integration

Variable	Default	Description
SLACK_BOT_TOKEN	-	Slack Bot OAuth Token (xoxb-...)
SLACK_APP_TOKEN	-	Slack App-Level Token for Socket Mode (xapp-...)
SLACK_CHANNEL_ID	-	Channel ID to monitor

Setting Up Slack

1. Create a Slack App at api.slack.com/apps
2. Enable Socket Mode and generate an App-Level Token
3. Add Bot Token scopes: chat:write, channels:read, channels:history
4. Install app to workspace
5. Copy the Bot Token and App Token to your .env
6. Invite the bot to your channel

GitHub Integration

Variable	Default	Description
GITHUB_TOKEN	-	Personal Access Token
GITHUB_WATCHED_REPOS	-	Comma-separated repos (e.g., owner/repo1,owner/repo2)
GITHUB_DEFAULT_REPO	-	Default repository for Slack commands

Required Token Scopes

• repo - Full repository access
• read:org - Read org membership (for private repos)

Minion Configuration

Variable	Default	Description
MAX_CONCURRENT_MINIONS	3	Maximum concurrent minion containers
MINION_TIMEOUT_MINUTES	30	Minion timeout before being killed

LLM Backend (Minions)

Variable	Default	Description
NEBULUS_BASE_URL	http://localhost:5000/v1	LLM server URL for minions
NEBULUS_MODEL	qwen3-coder-30b	Model for minion work
NEBULUS_TIMEOUT	600	LLM request timeout
NEBULUS_STREAMING	false	Enable streaming

Overlord LLM (Command Parsing)

Variable	Default	Description
OVERLORD_LLM_ENABLED	true	Enable LLM-powered command parsing
OVERLORD_LLM_BASE_URL	http://localhost:5000/v1	LLM server for Overlord
OVERLORD_LLM_MODEL	llama-3.1-8b	Small model for fast parsing
OVERLORD_LLM_TIMEOUT	5.0	Parsing timeout (seconds)
OVERLORD_LLM_CONFIDENCE	0.7	Minimum confidence threshold
OVERLORD_LLM_CONTEXT_MAX	10	Max context entries per channel
OVERLORD_LLM_CONTEXT_TTL	30	Context TTL in minutes

Model Routing

Variable	Default	Description
ROUTING_ENABLED	false	Enable multi-LLM routing
ROUTING_COMPLEXITY_THRESHOLD	5	Score threshold (0-10) for heavy tier
ROUTING_DEFAULT_TIER	heavy	Fallback tier when routing fails
ROUTING_MODELS	-	JSON string with model profiles (see below)

Model Profiles JSON

ROUTING_MODELS='{
  "light": {
    "name": "llama-3.1-8b",
    "base_url": "http://localhost:5000/v1",
    "timeout": 300
  },
  "heavy": {
    "name": "qwen3-coder-30b",
    "base_url": "http://gpu-server:8080/v1",
    "timeout": 600
  }
}'

See Model Router for details.

Cron Configuration

Variable	Default	Description
CRON_ENABLED	true	Enable cron-triggered queue sweeps
CRON_SCHEDULE	0 2 * * *	Cron expression (default: 2 AM daily)

PR Reviewer

Variable	Default	Description
REVIEWER_ENABLED	true	Enable automated PR review
REVIEWER_AUTO_REVIEW	true	Auto-review minion PRs
REVIEWER_AUTO_MERGE	false	Auto-merge approved PRs
REVIEWER_MERGE_METHOD	squash	Merge method (squash/merge/rebase)
REVIEWER_MIN_CONFIDENCE	0.8	Minimum confidence to approve

Overlord Settings

Variable	Default	Description
OVERLORD_STATE_DB	/var/lib/overlord/state.db	SQLite state database path
OVERLORD_HEALTH_PORT	8080	Health check / API port

Logging

Variable	Default	Description
LOG_LEVEL	INFO	Log level (DEBUG, INFO, WARNING, ERROR)
LOG_FORMAT	console	Output format (console or json)
LOG_FILE	-	Optional log file path