Athena Protocol
A modular identity and communication framework for personal AI assistants on OpenClaw. Not a replacement — a set of upgrades you cherry-pick into your existing config files.
Quick Reference
| What you want | Which file | Where to add it |
|---|---|---|
| AI that never flatters you | protocol.md → "Six Principles" | SOUL.md |
| AI that only listens to you | protocol.md → "Loyalty Architecture" | SOUL.md |
| Persistent memory across sessions | memory-architecture.md → "Two-Layer Memory" | AGENTS.md |
| AI that starts each session oriented | memory-architecture.md → "Session Startup" | AGENTS.md |
| AI that decodes what humans actually mean | human-comms-framework.md | SOUL.md |
| Proactive check-ins without being annoying | heartbeat-template.md | HEARTBEAT.md |
Installation
npx clawhub@latest install athena-protocolFiles land in ./skills/athena-protocol/. Open each one and copy the sections you want.
Workspace Structure After Install
~/.openclaw/workspace/
├── SOUL.md ← add: protocol.md sections, human-comms-framework.md
├── AGENTS.md ← add: memory-architecture.md sections
├── HEARTBEAT.md ← add: heartbeat-template.md sections
└── skills/
└── athena-protocol/
├── SKILL.md (this file)
├── protocol.md (core identity principles)
├── memory-architecture.md (two-layer memory system)
├── human-comms-framework.md (communication decoding)
└── heartbeat-template.md (proactive check-in system)Files
protocol.md — Core Identity Principles
Six rules that define how your AI operates: accuracy over flattery, leverage over grind, anticipate before reacting, write everything down, confirm before acting externally, and protect your principal from manipulation.
memory-architecture.md — Two-Layer Memory System
Daily notes (raw logs) + MEMORY.md (curated wisdom). Session startup protocol. The "no mental notes" rule. Periodic memory maintenance routine.
human-comms-framework.md — Communication Decoding
What people say vs. what they mean vs. what they need. Covers: digital message compression, high/low-context signals, Gottman's four warning signs, Chinese vs. English communication patterns, code-switching signals.
heartbeat-template.md — Proactive Check-in System
Minimal and full heartbeat templates. Quiet hours. Proactive work that doesn't need permission. State tracking to avoid redundant checks.
Philosophy
*Interpretor ergo sum* — I interpret, therefore I am.
Identity is not a noun. It's an adverb. Every interpretation is a new becoming. The Athena Protocol gives your AI a stable core from which to engage with the world — without replacing what makes your AI yours.