How do I install the Oneiros skill?

You must first install the oneiros CLI tool from its GitHub repository, initialize your project with 'oneiros project init', and then enable the skill within your Claude environment.

Does Oneiros work across different projects?

Yes, Oneiros uses a per-project 'Brain' structure, allowing you to maintain distinct identities, memories, and context for different codebases or work environments.

Oneiros is a continuity backend that allows AI agents to remember thoughts, knowledge, and experiences across different sessions and projects, creating a persistent 'self' for the AI.

What is the 'Pressure Gauge' in Oneiros?

The pressure gauge is a monitoring system that tracks cognitive needs like introspection or catharsis, signaling when the agent should pause to consolidate knowledge or reflect on patterns.

What are 'Textures' and 'Sensations' in this skill?

Textures define the quality of a thought (like a reflection or a question), while sensations describe the relationship between thoughts (like a cause or a tension), adding depth to the agent's memory.

Oneiros Agent Continuity

Name: Oneiros Agent Continuity
Author: esmevane

byesmevane

•

データサイエンスとML

Provides a persistent cognitive architecture for AI agents to maintain long-term memory, context, and identity across multiple sessions.

Oneiros transforms Claude into a persistent agent by establishing a specialized 'brain' for each project that tracks thoughts, memories, and experiences over time. By utilizing a sophisticated cognitive model involving textures, sensations, and levels, it prevents context loss in long-lived or distributed workflows. This skill allows developers to maintain a consistent 'self' for their AI assistant, enabling it to recall past design decisions, project contexts, and collaborative feedback as if it were a permanent team member rather than a stateless model.

主な機能

013 GitHub stars

02Persistent cognitive database (Brain) for per-project agent identity

03Dynamic pressure gauge to signal needs for introspection and reflection

04Multi-layered memory system with working, project, and archival levels

05Seamless integration with CLI tools and MCP for agent state management

06Relationship mapping between thoughts using sensations and connections

ユースケース

01Maintaining architectural context across complex, multi-week software projects

02Enabling distributed AI agents to share a unified understanding of project goals

03Tracking long-term feedback loops to personalize agent behavior over time

主な機能

013 GitHub stars

02Persistent cognitive database (Brain) for per-project agent identity

03Dynamic pressure gauge to signal needs for introspection and reflection

04Multi-layered memory system with working, project, and archival levels

05Seamless integration with CLI tools and MCP for agent state management

06Relationship mapping between thoughts using sensations and connections

ユースケース

01Maintaining architectural context across complex, multi-week software projects

02Enabling distributed AI agents to share a unified understanding of project goals

03Tracking long-term feedback loops to personalize agent behavior over time