Playbooks are KOMU's preferred operations path. They encode known workflows so the product does not depend only on provider-generated commands.
Playbooks are defined through a hybrid model:
- JSON catalog metadata
- Python renderer methods that produce typed plans
Each definition can include:
- category
- summary
- keywords and aliases
- metadata
- prerequisites
- platform and distro hints
- risk level
- dry-run support
- verification checks
- parameters
- repair playbooks
- optional rollback guidance
installconfigurediagnoserepairsecuritynetworkruntimeservice
- Docker install
- Nginx install
- Node.js install
- PM2 install
- Git install
- UFW setup
- Fail2ban setup
- SSL setup
- service restart
- service enable
- service status
- port diagnosis
- disk usage diagnosis
- process check
- log inspection
- system update
- package repair
- permission fix
- reverse proxy setup
The playbook flow is:
- score candidate playbooks from keywords and aliases
- extract parameters from the task
- apply defaults
- block the plan if required inputs are missing
- render typed steps and commands
- attach change summaries and rollback guidance where possible
- send the plan through the security review layer
Custom playbook catalogs can be added later through:
extensions/playbooks
This allows local or organization-specific playbooks to plug into the same match-render-review pipeline.