self-improving-operations

You capture operational learnings by logging incidents, toil, and capacity events into structured markdown entries within .learnings/ files. This creates a searchable knowledge base of reliability issues that teams can reference during future events.

The best way to turn recurring incident patterns into runbooks is by promoting repeated learnings from your .learnings/ files. Once a pattern surfaces multiple times, you elevate that entry into a formal runbook, postmortem, or SLO definition for repeatable reliability.

Yes, you can use markdown files to track operational issues and feature requests together. The Skill logs ongoing improvements to LEARNINGS.md, OPERATIONS_ISSUES.md, and FEATURE_REQUESTS.md, organizing both operational toil and automation gaps in one searchable structure.

Automation hooks integrate with templates to capture operational learnings automatically across multi-agent environments. These hooks support cross-skill references and standardized metadata, ensuring teams reliably extract and reuse knowledge without manual entry after every event.

You should promote operational learnings to capacity models or SLO definitions when recurring patterns emerge from logged incidents and toil. Recognizing repeated capacity events or reliability gaps signals that the learning needs elevation to a formalized operational standard.

Yes, this approach works for capturing toil and capacity events across multiple domains. It applies domain-agnostic logging to .learnings/ files and promotes recurring patterns to runbooks, automation backlogs, and SLO definitions, supporting cross-domain operational visibility.