One-click install
npx skills add https://github.com/Z5Jonathan-maker/ai-coding-os --skill consolidate-memory-z5jonathan-maker
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: consolidate-memory
Source: https://github.com/Z5Jonathan-maker/ai-coding-os/tree/main/claude/skills/consolidate-memory
Command: npx skills add https://github.com/Z5Jonathan-maker/ai-coding-os --skill consolidate-memory-z5jonathan-maker

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents valuable decisions, preferences, and facts from being lost across sessions by consolidating recent conversation logs into durable memory files.

Core Features & Use Cases

  • Recent session scanning: Reads recent session jsonl logs and extracts candidate facts from user/assistant conversational turns while skipping tool-result noise.
  • Memory de-duplication and conflict detection: Compares candidates against long-term, project-level, and wiki failure/optimization logs to avoid duplicates and flag contradictions for review rather than silently overwriting.
  • Memory layer updates: Appends new items into recent-memory, optionally promotes weekly recurring patterns into long-term memory, and refreshes project open threads/shipped lists in project-memory with per-project caps.
  • Use case: After several days of work on a project, running consolidation captures repeated preferences (e.g., “prefer minimal copy” or “use this repo structure”) and new facts (e.g., “installed X tool/dep”) so the next session starts smarter.

Quick Start

Run the consolidate-memory skill with the desired mode (nightly, weekly, or on-demand) and a time window to update recent, long-term (if weekly), and project memory from your local session logs.

Frequently Asked Questions about consolidate-memory

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

FAQPage Schema
How do I consolidate Claude session logs into a local knowledge base?

To consolidate Claude session logs into a local knowledge base, this Skill scans recent jsonl logs from ~/.claude/projects, extracts candidate facts, de-duplicates them against existing memory, and appends new entries into a 3-tier memory layer.

What is the best way to save project context and preferences across multiple sessions?

Saving project context across multiple sessions is achieved by distilling decisions and preferences from conversation logs into recent-memory, long-term-memory, and project-memory files, ensuring the next session starts with prior context.

How does memory de-duplication work when extracting facts from conversation logs?

Memory de-duplication works by comparing newly extracted candidate facts against existing long-term and project-level memory logs to avoid duplicates and actively flag contradictions for review rather than silently overwriting old data.

Can I run memory consolidation on a nightly or weekly schedule for multi-session projects?

Yes, you can run memory consolidation on a nightly, weekly, or on-demand schedule. Weekly runs optionally promote recurring patterns into long-term memory, while nightly runs refresh recent memory and active project threads with per-project caps.

Does memory consolidation automatically extract preferences from jsonl session files?

Yes, memory consolidation automatically extracts preferences by reading recent session jsonl logs and scanning user and assistant conversational turns while skipping tool-result noise to identify repeated preferences and new project facts.

What are the limitations of consolidating memory from local session logs?

A limitation of consolidating memory from local session logs is that processing is restricted to jsonl files under ~/.claude/projects, and extracted facts are subject to per-project caps for open threads and recently shipped lists.