One-click install
npx skills add https://github.com/onebrain-ai/onebrain --skill recap-onebrain-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: recap
Source: https://github.com/onebrain-ai/onebrain/tree/main/.claude/plugins/onebrain/skills/recap
Command: npx skills add https://github.com/onebrain-ai/onebrain --skill recap-onebrain-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Recap turns repeated lessons from your recent session logs into long-term memory files, so patterns you keep returning to become persistent knowledge instead of getting lost after each session.

Core Features & Use Cases

  • Frequency-filtered promotion: Promotes only insights that appear across multiple separate session logs, avoiding one-off noise.
  • Conflict-aware consolidation: Detects conflicts against existing active/needs-review memory files and asks how to handle them (update, supersede, separate, or skip).
  • Memory consolidation: Merges overlapping memory files when appropriate, using conservative confidence rules and respecting contradictions.
  • Session log bookkeeping: Marks processed logs as recapped, derives topics/keywords, and writes a per-day audit log entry.

Quick Start

Run the recap skill to extract and promote recurring insights from your recent unrecapped session logs, using your onebrain.yml frequency thresholds.

Frequently Asked Questions about recap

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I turn recurring session insights into persistent memory files?

Yes, frequency filtering prevents one-off observations from polluting your knowledge base by applying threshold rules from a configuration file to ensure insights appear across multiple separate session logs before promotion.

How does conflict resolution work when consolidating overlapping memory topics?

Conflict resolution detects contradictions against existing active memory files and interactively asks how to handle them, allowing you to update, supersede, separate, or skip conflicting knowledge entries during consolidation.

What is the best way to batch process unrecapped session logs for knowledge extraction?

The best way to batch process unrecapped session logs is using glob-based discovery to find recent entries, extract recurring insights, resolve memory conflicts, and mark processed logs as recapped with updated frontmatter and stats.

How do I keep an audit log of memory promotion updates from session notes?

If one-off observations pollute your knowledge base, you need frequency filtering to require that insights appear across multiple separate session logs before they are promoted into persistent memory files.