Continual Learning & Memory FAQs

Question 1

Where is the learned information saved?

Accepted Answer

Extracted learnings are formatted as plain bullet points and stored in an AGENTS.md file within your repository, typically managed on a dedicated branch to ensure persistence across sessions.

Question 2

Does it process the entire history every time I run it?

Accepted Answer

No, it utilizes an incremental indexing system that tracks file modification times, ensuring that only new or updated transcripts are processed to maintain high performance.

Question 3

Does this skill store sensitive information like API keys?

Accepted Answer

No. The skill features a strict inclusion bar and exclusion list that explicitly filters out secrets, credentials, tokens, and transient data, focusing only on high-signal architectural patterns and preferences.

Question 4

How does it handle memory across different git branches?

Accepted Answer

The skill fetches the latest remote state and uses a centralized branch strategy (defaulting to a 'continual-learning' branch) to ensure your AI's memory is synchronized across the entire project regardless of your current local branch.

Question 5

How does the skill discover my chat transcripts?

Accepted Answer

It automatically scans standard directories for both Cursor and Claude Code, using git worktree discovery to map project paths to their respective transcript storage locations across all repository checkouts.

Continual Learning & Memory

주요 기능

사용 사례

Continual Learning & Memory

주요 기능

사용 사례