Create new peer-level AI agents with customized personalities and independent workspaces using a secure, gated replication process.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install bobiverse-replicate
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 bobiverse-replicate 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 Replicate: Purposeful Agent Cloning?
Replicate is a sophisticated cloning skill for AI agents, inspired by the purposeful replication concepts found in the Bobiverse. Unlike simple sub-agent spawning, this skill creates top-level peer agents with their own distinct workspaces, identities, and memory stores. By utilizing this addition to Openclaw Skills, developers can scale their agentic operations while maintaining strict operator control and security boundaries.
At its core, Replicate focuses on intentionality. It is designed to prevent runaway replication by enforcing hard gates, requiring clear mission statements, and utilizing a hardened Python runner. This ensures that every new agent added to the Openclaw Skills ecosystem has a specific purpose and a verified lineage.
Replicate: Purposeful Agent Cloning Use Cases
- Explicit operator invocation for task-specific scaling
- Specializing a clone for a new project or star system domain
- Experimenting with personality variations and SOUL.md modifications
- Pruning memory for a fresh start on complex task boundaries
How Replicate: Purposeful Agent Cloning Works
- Trigger and Validation: The operator issues an explicit command and provides a concrete mission statement.
- Parameter Gathering: The agent collects inputs for the clone name, star system, personality tweaks, and memory policy.
- Safe Dry-Run: A hardened script (replicate_safe.py) generates a replication plan and a unique, one-time confirmation token.
- Operator Approval: The operator provides the nonce-backed confirmation token (e.g., REPLICATE
). - Execution: The runner performs a transactional workspace duplication within the Openclaw directory tree.
- Post-Clone Customization: The skill updates the clone's SOUL.md, IDENTITY.md, and MEMORY.md based on the provided parameters.
- Registration and Reporting: The new agent is registered via the CLI and added to the parent's LINEAGE.md tracking file.
Replicate: Purposeful Agent Cloning Setup
To install the Replicate skill via the official registry, use the following command:
openclaw skills install agent-replicate
Ensure that python3 is available in your PATH, as the secure runner requires it for validation and file operations. For manual installations, ensure the bundled assets like SERIAL-NUMBER-SPEC.md and the scripts/ directory are located within your skill path.
Replicate: Purposeful Agent Cloning Data Schema & Taxonomy
The replication process organizes data through several core files within the new agent's workspace:
| File | Description |
|---|---|
| IDENTITY.md | Contains the serial number, generation, and parent lineage. |
| SOUL.md | Defines the clone's personality and includes Divergence Notes. |
| MEMORY.md | Stores inherited or pruned knowledge based on the chosen memory policy. |
| LINEAGE.md | A runtime ledger documenting the fork tree and agent registry. |
Replicate: Purposeful Agent Cloning Advanced Features
- Nonce-Backed Confirmation: Prevents accidental replication by requiring a specific one-time token from a dry-run output.
- Granular Memory Policies: Choose between full inheritance, pruned observations, or minimal baseline memory.
- Personality Divergence: Automatically generates a Divergence Notes section in SOUL.md to track trait variations.
- Cross-Agent Communication: Automated setup for parent-clone messaging within the Openclaw Skills framework.
- Hardened Security Runner: Rejects symlinks and enforces path boundaries to prevent filesystem escapes.
SKILL.md
Loading
Related Openclaw Skills
Benchmark Store
A specialized performance management engine that tracks historical baselines and prevents multi-dimensional regressions in AI agents.
Monthly Weekend Travel Blind Box
An AI-driven travel recommendation engine that delivers weather-optimized weekend trip blind boxes complete with real-time transportation and booking details.
Enterprise Training Outline Creator
A professional tool for generating structured corporate training outlines and exporting them as formatted DOCX documents.
A3 PDF Smart Splitter
An intelligent utility that automatically splits A3-sized PDF exam papers into A4 pages by detecting the optimal white space for a clean cut.
ZAP1 - Zcash Attestation
An attestation layer that anchors AI agent actions and policy decisions to the Zcash mainnet using cryptographic Merkle tree proofs.
TikTok Trend Intelligence
A real-time intelligence engine for TikTok that monitors trending hashtags, viral music, and high-growth products to identify market opportunities.
METADATA
Featured*
