Safety-first DevOps preset pack for Clawdbot
Pre-configured safety presets and DevOps skills for using Clawdbot as a DevOps assistant. Get a ChatOps interface to your infrastructure through Telegram, Slack, or terminal.
- Safety-First Design - Three safety modes (safe, standard, full)
- DevOps Skills - Kubernetes, Docker, Terraform, AWS, and more
- Multiple Channels - Telegram, Slack, Discord, or TUI
- Uses Your Existing Auth - Works with your Anthropic subscription via Clawdbot
# One-liner setup
curl -fsSL https://raw.githubusercontent.com/agenticdevops/moltbot-devops-presets/main/scripts/setup.sh | bashOr step by step:
# 1. Clone this repo
git clone https://github.com/agenticdevops/moltbot-devops-presets.git
cd moltbot-devops-presets
# 2. Run setup
./scripts/setup.sh
# 3. Start clawdbot
clawdbot| Mode | Description | When to Use |
|---|---|---|
| Safe | Read-only operations only | Production monitoring |
| Standard | Mutations require approval | Day-to-day operations |
| Full | All operations allowed | Development/testing only |
Only allows read operations (get, describe, list, logs). Any mutation is blocked.
Read operations auto-execute. Mutations prompt for approval:
> Scale the api deployment to 5 replicas
🔸 Command requires approval:
kubectl scale deployment/api --replicas=5
[Allow] [Deny] [Allow All Similar]
All commands execute without approval. Use with caution!
Includes 11 DevOps skills from agentic-ops-skills:
| Skill | Description |
|---|---|
k8s-deploy |
Safe Kubernetes deployments and rollbacks |
k8s-debug |
Kubernetes debugging and troubleshooting |
docker-ops |
Docker container management |
aws-ops |
AWS operations (EC2, S3, Lambda, ECS) |
terraform-workflow |
Infrastructure as Code workflows |
incident-response |
Structured incident response |
system-health |
System monitoring and diagnostics |
log-analysis |
Cross-platform log analysis |
argocd-gitops |
GitOps with ArgoCD |
cost-optimization |
Cloud cost management |
git-workflow |
DevOps git practices |
> Check my cluster health
> What pods are failing in production?
> Show me the logs for the api pod
> Scale the frontend deployment to 3 replicas
> List running containers
> Show me the nginx container logs
> What's using the most resources?
> Plan the changes for the staging environment
> Show me the current state
> What resources would be destroyed?
> We have high latency on the API. Help me investigate.
> Check the error rate for the payment service
> Create an incident summary
- Create a bot via @BotFather
- Set the token:
export TELEGRAM_BOT_TOKEN=your-token - Start clawdbot:
clawdbot
- Create a Slack app at https://api.slack.com/apps
- Configure Socket Mode and Bot Token
- Set credentials:
export SLACK_BOT_TOKEN=xoxb-... export SLACK_APP_TOKEN=xapp-...
Just run clawdbot - the TUI is always available.
Config is stored at ~/.clawdbot/clawdbot.json. You can:
- Switch presets by copying from
presets/ - Modify allowlist patterns
- Enable/disable specific skills
- Configure channel settings
# Switch to safe mode
cp presets/devops-safe.json ~/.clawdbot/clawdbot.json
# Switch to standard mode
cp presets/devops-standard.json ~/.clawdbot/clawdbot.json- Node.js 18+ or Bun
- Git
- Clawdbot (installed automatically by setup)
- Anthropic API key (via Clawdbot auth)
┌─────────────────────────────────────────────────────────┐
│ Moltbot DevOps Presets │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │
│ │ Presets │ │ Skills │ │ Setup │ │
│ │ (JSON) │ │ (Symlinks) │ │ Script │ │
│ └──────┬──────┘ └──────┬──────┘ └─────────────┘ │
└─────────┼────────────────┼──────────────────────────────┘
│ │
▼ ▼
┌─────────────────────────────────────────────────────────┐
│ Clawdbot │
│ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ │
│ │ Agent │ │ Channels│ │ Exec │ │ Auth │ │
│ │ Runtime │ │ TG/Slack│ │Approvals│ │Anthropic│ │
│ └─────────┘ └─────────┘ └─────────┘ └─────────┘ │
└─────────────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ agentic-ops-skills │
│ k8s-deploy, docker-ops, aws-ops, terraform, ... │
└─────────────────────────────────────────────────────────┘
- Fork the repository
- Add/modify presets in
presets/ - Submit a PR
For skill contributions, see agentic-ops-skills.
MIT