patterns

Skill: Patterns

Purpose

Browse and search recurring patterns discovered across analyses. Patterns are auto-extracted after each analysis archive and represent behaviors that appear consistently in the data.

When to Use

• User says /patterns or "what patterns have we seen?"
• During analysis, to check if a finding matches a known pattern
• At session start, to remind the user of established behaviors

Invocation

/patterns — list patterns for the active dataset /patterns --global — list patterns across all datasets /patterns search={term} — search patterns by keyword /patterns {id} — show full details for a specific pattern

Instructions

Step 0: Determine Active Dataset

Before loading patterns, identify the active dataset:

1. Read .knowledge/active.yaml to get the active dataset name
2. If the file doesn't exist or is empty, default to checking all datasets
3. Use this dataset name when filtering patterns and referencing dataset-specific files

This ensures you're searching patterns for the correct dataset and providing accurate context.

Step 1: Load Patterns

1. Check if .knowledge/analyses/_patterns.yaml exists:
   • If it doesn't exist: "No patterns recorded yet. The pattern system initializes after your first analysis is archived."
   • If it exists but is empty: "No patterns recorded yet. Complete 2-3 analyses and recurring patterns will emerge."
2. If --global flag: also check and read .knowledge/global/cross_dataset_observations.yaml (same existence checks apply).
3. Load pattern data from existing files.

Step 2: Execute Command

List patterns (/patterns):

• Filter to active dataset (unless --global)
• Sort by occurrences descending (most established first)
• Display as a table: type, description, occurrences, confidence, last seen
• Show total count

Show specific (/patterns {id}):

• Display: description, type, all evidence (with analysis IDs), dimensions, metrics, suggested investigation
• Offer: "Want to investigate this pattern further?"