Masaomi Kairoschain 2026 Kairos Chain

KairosChain - Self-Amendment MCP Server Framework

KairosChain provides a layered skill architecture (L0/L1/L2) where skills can evolve, promote, and audit themselves with blockchain-backed auditability.

Architecture

Three-Layer System

• L0 (Constitution/Law): Immutable safety rules and meta-governance. Changes require human approval and full blockchain recording.
• L1 (Knowledge): Project knowledge in Anthropic skills format. Changes recorded with hash references.
• L2 (Context): Temporary session context. Free modification, no blockchain recording.

Core Capabilities

Knowledge Management

• knowledge_list / knowledge_get - Browse and read L1 knowledge skills
• knowledge_update - Create, update, or delete L1 knowledge with blockchain recording
• context_save - Save temporary L2 context for session work

Skill Evolution

• skills_evolve - Propose and apply changes to L0 skill definitions (requires human approval)
• skills_rollback - Version management with snapshot and rollback capabilities
• skills_promote - Promote knowledge between layers (L2→L1, L1→L0) with optional Persona Assembly

Blockchain Auditability

• chain_status / chain_verify - Check and verify blockchain integrity
• chain_history - View skill transitions, knowledge updates, and state commits
• chain_record - Record data to the blockchain
• state_commit - Create snapshots of all layers for auditability

Health & Safety

• skills_audit - Audit knowledge health across layers (conflicts, staleness, dangerous patterns)
• tool_guide - Dynamic tool discovery and workflow recommendations

Persona Assembly

• Multi-perspective decision support using personas (kairos, conservative, radical, pragmatic, optimistic, skeptic)
• Available in skills_promote and skills_audit with assembly_mode: "discussion"

Recommended Workflows

Knowledge Lifecycle

1. Start with context_save for temporary work (L2)
2. When knowledge proves valuable, use skills_promote to move L2→L1
3. For mature patterns, promote L1→L0 with human approval

Health Check

1. Run skills_audit command="check" for overall health
2. Use skills_audit command="dangerous" to check for safety issues
3. Review skills_audit command="recommend" for promotion/archive suggestions

Skill Evolution

1. Use skills_dsl_list / skills_dsl_get to inspect current L0 skills
2. Propose changes with skills_evolve command="propose"
3. Apply with human approval via skills_evolve command="apply"

Converted and distributed by TomeVault — claim your Tome and manage your conversions.