Linux Ops Skill v2.0
This is an Agent-native Linux operations runtime over SSH.
It is not an Ansible clone, not a hidden playbook runner, and not a pile of app-specific repair scripts. The Agent uses Linux and operations knowledge to reason from live observations; the skill provides bounded primitives, strict contracts, fail-closed runtime gates, policy validation, redaction, audit, and escalation.
Recommended loop:
observe -> classify -> hypothesize -> write decision record ->
[Description truncada. Veja o README completo no GitHub.]