Does this skill require an AWS account?

Yes, you must have an active AWS account with Amazon Bedrock AgentCore permissions, an agent deployed, and the necessary IAM roles configured.

What is the difference between episodic and short-term memory?

Short-term memory tracks the raw conversation context of the current session, while episodic memory stores extracted long-term facts and preferences that persist across different sessions.

Can I manually add or delete specific memories?

Yes, the skill provides specific operations for manual memory management, allowing you to create, update, or delete records independently of the automatic extraction process.

How does the agent 'learn' new facts?

It uses a reflection engine that processes session data in the background or at the end of a session to extract significant facts and store them in episodic memory.

Is user data kept private between different actors?

Yes, the memory architecture is user-scoped (actor-scoped), meaning knowledge is isolated per user ID to prevent data leakage between different users.

Amazon Bedrock Agent Memory

Name: Amazon Bedrock Agent Memory
Author: adaptationio

byadaptationio

クラウドインフラストラクチャ

Integrates Amazon Bedrock AgentCore to provide persistent episodic and short-term memory for intelligent agents across multiple sessions.

概要

Amazon Bedrock AgentCore Memory empowers your agents to evolve from stateless bots into intelligent assistants that remember user preferences and learn from past interactions. By combining raw short-term session context with a sophisticated reflection engine that extracts long-term episodic facts, this skill enables a highly personalized user experience. It handles semantic retrieval for context-aware lookups and maintains strict user-scoped privacy, making it ideal for building sophisticated customer service, travel, or productivity agents on the AWS ecosystem.

主な機能

High-performance semantic memory retrieval (<100ms)
Dual-layer episodic and short-term memory architecture
Granular user-scoped data isolation
Manual memory record management and batch operations
Automatic background reflection for fact extraction
0 GitHub stars

ユースケース

Personalized assistants that remember user preferences and historical choices
Technical support agents that maintain context across multiple service tickets
Educational tutors that track and adapt to a learner's progress over time

概要

主な機能

High-performance semantic memory retrieval (<100ms)
Dual-layer episodic and short-term memory architecture
Granular user-scoped data isolation
Manual memory record management and batch operations
Automatic background reflection for fact extraction
0 GitHub stars

ユースケース

Personalized assistants that remember user preferences and historical choices
Technical support agents that maintain context across multiple service tickets
Educational tutors that track and adapt to a learner's progress over time