How does this skill help save tokens during search?

It uses a tiered approach where it first performs a lightweight search of IDs and titles (~50 tokens per result) and only fetches full details for the specific items you select.

What kind of information should I store in this memory system?

It is best used for recording architectural decisions, recurring error solutions, project-specific conventions, and high-level summaries of what was accomplished in a session.

Does this work with other AI editors besides Claude Code?

Yes, while this is optimized as a Claude Code Skill, the underlying AgentKits Memory system is compatible with Cursor, Copilot, Windsurf, and Cline via MCP.

Can I filter my search by date or category?

Absolutely. You can filter results by categories like 'decision' or 'error', and provide specific date ranges to find exactly when a change was discussed.

AgentKits Persistent Memory

Name: AgentKits Persistent Memory
Author: aitytech

byaitytech

•

Productivity & Workflow

Maintains long-term project context and historical decision logs using a token-efficient three-layer retrieval system.

AgentKits Memory provides Claude with a persistent database of past work, architectural decisions, and error resolutions that spans across multiple coding sessions. By implementing a specialized three-layer search workflow—ranging from lightweight indexing to deep detail retrieval—the skill ensures that the assistant can recall specific project history while reducing token consumption by up to 87%. It is an essential tool for developers working on complex, long-term projects who require their AI assistant to maintain a consistent 'mental model' of the codebase's evolution.

Key Features

01Categorized memory storage for decisions, patterns, errors, and observations

02Database status monitoring to verify available context before searching

0352 GitHub stars

04Quick topic recall for immediate summaries of specific project areas

05Three-layer search workflow for high-precision, token-efficient retrieval

06Timeline context viewing to understand the sequence of historical events

Use Cases

01Recalling the rationale behind specific architectural patterns or past technical decisions

02Maintaining continuity across different development branches and long-running feature implementations

03Retrieving previous solutions to recurring bugs or configuration errors encountered in earlier sessions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add aitytech/agentkits-memory memory

For use in Claude.ai and ChatGPT

Download Skill