Compound Learnings

This is the meta-skill that turns your Claude sessions into permanent infrastructure. It reads through your learnings cache, extracts patterns that show up across multiple sessions, and transforms them into rules, skills, or hooks based on what they actually are. The decision tree is smart: sequences become executable skills, heuristics become rules, recurring tasks become hooks. Most interesting part is the consolidation step where it merges similar patterns before counting frequency, so "artifact-first debugging" and "filesystem-first debugging" get recognized as the same principle. Run it when you notice yourself repeatedly learning the same lessons, or when a topic keeps appearing and you suspect it needs dedicated tooling rather than scattered rules.

skills.sh installs: 507
GitHub stars: 3.9K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add parcadei/continuous-claude-v3 --skill compound-learnings --agent claude-code

Installs into .claude/skills of the current project.

Files

Compound Learnings

Transform ephemeral session learnings into permanent, compounding capabilities.

When to Use

"What should I learn from recent sessions?"
"Improve my setup based on recent work"
"Turn learnings into skills/rules"
"What patterns should become permanent?"
"Compound my learnings"

Process

Step 1: Gather Learnings

# List learnings (most recent first)
ls -t $CLAUDE_PROJECT_DIR/.claude/cache/learnings/*.md | head -20

# Count total
ls $CLAUDE_PROJECT_DIR/.claude/cache/learnings/*.md | wc -l

Read the most recent 5-10 files (or specify a date range).

Step 2: Extract Patterns (Structured)

For each learnings file, extract entries from these specific sections:

Section Header	What to Extract
`## Patterns` or `Reusable techniques`	Direct candidates for rules
`Takeaway:` or `Actionable takeaway:`	Decision heuristics
`## What Worked`	Success patterns
`## What Failed`	Anti-patterns (invert to rules)
`## Key Decisions`	Design principles