The problem it solves
Generic AI tools don't know your team's patterns: how you structure staging models, name CTEs, order columns, or format YAML. They have to learn them somehow.
What it does
Reads a good example file from the codebase.
Extracts reusable patterns rather than copying content.
Asks for confirmation before saving.
Stores the pattern as training for future sessions.
When to use it
WHEN
onboarding Altimate Code to a project
WHEN
a file demonstrates the style the team wants repeated
WHEN
replacing long instructions with concrete examples
WHEN
team conventions are implicit rather than documented
How it works
The skill reads the file, pulls out the structural patterns, shows you what it found for confirmation, and saves them so future sessions can apply them.
Tools used
training_save
Save learned patterns, rules, glossary terms, standards, context, or playbooks.
training_list
List saved training entries with metadata and usage information.