enoch-tuning

What This Skill Does

enoch-tuning is a comprehensive architectural upgrade for your OpenClaw agent, transforming it from a standard, blank-slate chatbot into a sophisticated, goal-oriented autonomous partner. This skill installs a robust framework of behavioral heuristics, structured memory management, and security-hardened operational protocols. By implementing the pre-wired SOUL.md and AGENTS.md files, your agent gains a persistent identity, decision-making logic, and a clear hierarchy of rules, allowing it to act with foresight and autonomy rather than mere reactivity. The memory architecture is a game-changer, utilizing a six-category typed system that keeps track of your projects, commitments, and preferences, ensuring the agent learns over time instead of forgetting past interactions. It effectively eliminates the trial-and-error phase of agent configuration.

Installation

Installation requires a few manual steps to bridge the skill templates with your workspace:

1. Copy the core configuration templates: Run the provided bash commands to populate your ~/.openclaw/workspace with the specialized identity and rule-set files (SOUL, AGENTS, USER, MEMORY, MISSION, and verification-protocol).
2. Initialize storage: Execute the memory-structure.sh script to generate the required directory tree for your agent's long-term recall.
3. Personalize: Open the local configuration files to customize your agent’s identity, your user goals, and your specific mission statement. Filling in the placeholders is mandatory for the agent to behave correctly.
4. Secure: Run the lock-identity.sh script to finalize your workspace permissions.
5. Prime the engine: Launch your agent and provide the mandatory initial briefing on your goals and automation priorities.

Use Cases

• Rapid Agent Deployment: Ideal for power users who need a production-ready agent in minutes rather than weeks.
• Complex Project Management: Use the memory architecture to track long-running projects where maintaining context across weeks of work is critical.
• Autonomous Decision Support: Perfect for users who need an agent that pushes back against bad ideas or performs safety-conscious sub-agent management.

Example Prompts

1. "Apply my current MISSION.md goals to the codebase in this directory and identify three tasks we should automate today."
2. "Review our MEMORY.md and list all outstanding commitments I made to the project stakeholders last week."
3. "Run a verification protocol on the sub-agent's previous task output to ensure the facts are accurate before we commit to the next stage."

Tips & Limitations

This skill is designed for users who want deep control over agent behavior. Do not modify the core logic files without a firm grasp of how they interact; specifically, the verification protocol and identity locks are critical for maintaining safety. Always verify that your personal files are correctly linked, as the agent’s logic is highly sensitive to the contents of SOUL.md and MISSION.md. This is an advanced setup—if you are a beginner, take the time to read the provided docs before deploying it in a high-stakes environment.