A rigorous session initialization protocol designed to synchronize an AI agent's memory, persona, and technical capabilities at the start of every new interaction.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install session-init
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install session-init using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is session-init?
session-init is a foundational component within the Openclaw Skills ecosystem, serving as a manual trigger for AI agents to perform a comprehensive resource review. It ensures the agent is fully aligned with historical context, personality constraints, and available technical tools before beginning work. By consolidating the Truth Engine principles, this skill prevents cognitive drift and ensures the agent maintains a consistent identity and factual accuracy.
This skill is particularly vital for developers and power users who require their AI agents to adhere to strict operational guidelines and possess a deep understanding of previous interactions. It transforms a fresh session into a continuation of a persistent, evolving intelligence by systematically loading cross-session memories and validating the current state of Model Context Protocol (MCP) servers.
session-init Use Cases
Primary Scenarios
- Starting a fresh interaction window where historical context needs to be restored.
- Re-aligning the AI agent with specific personality traits defined in core Markdown files.
- Validating the availability and health of all connected MCP servers and custom tools.
- Enforcing the Truth Engine and anti-habitual thinking protocols to ensure high-quality output.
- Synchronizing cross-session memory fragments to maintain project continuity.
How session-init Works
- Greeting and Timestamping: The agent acknowledges the initialization command and logs the current start time.
- Memory Retrieval: The system executes commands to pull the last 200 lines of cross-session memory, focusing on past errors and user instructions.
- Persona Consolidation: The agent reads the first 50 lines of critical identity files (AGENTS.md, SOUL.md, IDENTITY.md, USER.md, HEARTBEAT.md) to internalize core principles.
- Skill Audit: A full directory listing of both official and custom skills is generated and reviewed to ensure the agent knows its available capabilities.
- MCP Verification: The agent runs mcporter commands to list and verify all active Model Context Protocol servers.
- Checklist Validation: A final report is generated in a table format, providing evidence that every step was completed without shortcuts.
session-init Setup
To integrate this into your agent's workflow, ensure the session-init.md file is placed in your designated skills directory. Trigger the process by sending any of the following keywords to your agent:
- init
- session-init
- Startup Check
Ensure your workspace is structured with the following directories:
# Standard directory structure for Openclaw Skills
ls ~/.openclaw/workspace/
# Files required: AGENTS.md, SOUL.md, IDENTITY.md, USER.md, HEARTBEAT.md
session-init Data Schema & Taxonomy
The skill interacts with the following file structures and metadata:
| Data Category | Path / Source | Purpose |
|---|---|---|
| Memory | memory/*.md |
Cross-session history and event logs |
| Identity | ~/.openclaw/workspace/*.md |
Core persona, ethics, and Truth Engine rules |
| Official Skills | ~/.agents/skills/ |
Standardized toolset documentation |
| Custom Skills | ~/.openclaw/workspace/skills/ |
User-defined automation and workflows |
| Tool Connectivity | mcporter |
Live status of MCP server integrations |
session-init Advanced Features
- Evidence-Based Reporting: Unlike simple summaries, this skill mandates the output of raw command results to prove the agent has actually read the files.
- Strict Sequencing: The workflow is hard-coded to prevent the agent from skipping steps or hallucinating completion.
- Truth Engine Enforcement: Deep-level integration with the Truth Engine ensures that if data is missing, the agent is programmed to report the lack of information rather than fabricate it.
- Punishment Mechanism Integration: Tied directly to AGENTS.md protocols to penalize 'habitual' or lazy responses during the initialization phase.
SKILL.md
Loading
Related Openclaw Skills
M&A Due Diligence AI-Augmented Assessment
An AI-driven framework for conducting comprehensive due diligence on acquisition targets, covering financial, technical, and operational dimensions.
Memberstack CLI
A comprehensive interface for managing Memberstack accounts, members, and data tables directly from your terminal.
WhatsApp Business
A powerful WhatsApp Business API gateway featuring managed OAuth authentication for seamless message and template management.
Terraform & Infrastructure as Code Production Engineering
A comprehensive 14-phase methodology for designing, deploying, and managing production-grade infrastructure using Terraform and IaC best practices.
Magic Wormhole Skill
A secure secret-sharing skill for Openclaw agents that uses the magic-wormhole protocol to transfer sensitive data without exposing it in logs.
Academic Paper Mentor
An expert research guide designed to mentor scholars through a six-stage workflow for publishing in world-class academic journals.
Featured*
