enoch-tuning

Most people who set up an AI agent get a blank slate. No memory, no personality, no rules. They spend weeks figuring out why it keeps forgetting things, why it sounds like a chatbot, why it won't push back when they're wrong.

This skill skips all of that.

What you're installing is a production-tested identity and memory system — decision heuristics, hard rules, security protocols, memory architecture, and automation pipelines that took months to develop and refine.

What You Get

• Pre-wired SOUL.md — decision heuristics, hard rules, anti-patterns, cost awareness. The behavioral core that makes the difference between a useful agent and a corporate chatbot.
• AGENTS.md — full operating rules: verification protocol, status reporting, Claude Code coordination, AFK behavior, sub-agent management, safety tiers, idiot prevention.
• Memory architecture — 6-category typed memory system (decisions, people, lessons, commitments, preferences, projects), VAULT_INDEX, daily log structure.
• MISSION.md template — mission-driven idle behavior. Agent asks "what gets us closer to the mission?" instead of waiting.
• Verification protocol — prevents stale data, fake sub-agent completions, and unverified facts from reaching you.
• Setup scripts — memory directory structure, identity file locking.

Installation

Step 1 — Copy templates

cp skills/enoch-tuning/templates/SOUL.md ~/.openclaw/workspace/SOUL.md
cp skills/enoch-tuning/templates/AGENTS.md ~/.openclaw/workspace/AGENTS.md
cp skills/enoch-tuning/templates/USER.md ~/.openclaw/workspace/USER.md
cp skills/enoch-tuning/templates/MEMORY.md ~/.openclaw/workspace/MEMORY.md
cp skills/enoch-tuning/templates/MISSION.md ~/.openclaw/workspace/MISSION.md
cp skills/enoch-tuning/templates/ops/verification-protocol.md ~/.openclaw/workspace/ops/verification-protocol.md

Step 2 — Create memory structure

bash skills/enoch-tuning/setup/memory-structure.sh ~/.openclaw/workspace

Step 3 — Personalize (required)

Edit these files — everything in [BRACKETS] is a placeholder:

• SOUL.md — name, worldview, vibe
• USER.md — your info, goals, rhythm
• MEMORY.md — your platform setup, key facts
• MISSION.md — your mission statement (one sentence)

Step 4 — Lock identity files

bash skills/enoch-tuning/setup/lock-identity.sh ~/.openclaw/workspace

Step 5 — First conversation

Tell your agent: your name, what you do, the top 3 things you want automated, and what it should never do without asking. Everything compounds from here.

What NOT to Change Without Understanding

• Hard Rules section in SOUL.md — these are non-negotiable behavioral guardrails
• Idiot Prevention Protocol in AGENTS.md — protects your infrastructure from chat-based config changes
• Verification Protocol — removing this reintroduces stale data and fake completions
• Automation tiers — the boundary between "runs without asking" and "never without instruction" is load-bearing

File Structure

skills/enoch-tuning/
├── SKILL.md                          ← this file
├── templates/
│   ├── SOUL.md                       ← identity template
│   ├── AGENTS.md                     ← operating rules template
│   ├── USER.md                       ← user intake template
│   ├── MEMORY.md                     ← long-term memory template
│   ├── MISSION.md                    ← mission statement template
│   └── ops/
│       └── verification-protocol.md  ← fact-checking protocol
└── setup/
    ├── memory-structure.sh           ← creates memory directories
    └── lock-identity.sh              ← locks SOUL.md + AGENTS.md