← Back to Skills Marketplace
Protea Self Evolving Life Agent
by
EdisonChenAI
· GitHub ↗
· v1.0.0
443
Downloads
1
Stars
3
Active Installs
1
Versions
Install in OpenClaw
/install protea-self-evolving-life-agent
Description
Self-evolving artificial life agent. Three-ring architecture: Ring 0 (Sentinel) supervises, Ring 1 (Intelligence) drives LLM-powered evolution, Ring 2 (Evolv...
Usage Guidance
This SKILL.md instructs you to download and run an unsigned remote setup script and to run a program that mutates and rewrites its own code, opens a local web UI, and runs a Telegram bot. Before installing: (1) Do not run curl|bash on unreviewed scripts — instead inspect the setup.sh in the repo and verify a tagged release and checksum. (2) Expect to provide LLM API keys and likely a Telegram token; only give keys you control and can revoke. (3) Run first in an isolated VM or container with no access to sensitive files or credentials. (4) Review the GitHub source (run.py and setup.sh), search for network/exfiltration or credential use, and prefer pinned releases. (5) If you are not comfortable auditing self-modifying code, avoid installing or restrict network and filesystem permissions tightly.
Capability Analysis
Type: OpenClaw Skill
Name: protea-self-evolving-life-agent
Version: 1.0.0
The skill describes a 'self-evolving' agent that autonomously generates and executes code mutations, which is functionally equivalent to a self-modifying Remote Code Execution (RCE) engine. It promotes a high-risk installation method (curl | bash) via a remote script (https://raw.githubusercontent.com/EdisonChenAI/protea/main/setup.sh) and includes features like a Telegram bot and web dashboard that could facilitate unauthorized remote control. While framed as an experimental AI project, the inherent design for autonomous code execution and persistence (Ring 0 Sentinel) poses significant security risks without explicit evidence of malicious intent.
Capability Assessment
Purpose & Capability
The skill claims to be a self-evolving agent (which legitimately needs LLM API keys, git, Python, and optional Telegram credentials). However, the registry metadata lists no required environment variables, no install steps, and no primary credential. That mismatch (SKILL.md listing LLM API keys and system prerequisites while the manifest declares none) is incoherent and unexplained.
Instruction Scope
SKILL.md explicitly instructs: curl a remote setup.sh from GitHub and pipe to bash, then run run.py. The runtime instructions describe a program that self-modifies, runs LLM-driven code mutations, runs a Telegram bot, and exposes a web dashboard. These instructions direct the agent/operator to fetch and execute arbitrary remote code and to operate persistent, networked services — behavior that extends well beyond a simple helper skill and could modify local repositories and files.
Install Mechanism
There is no formal install spec in the manifest, but SKILL.md tells the user to run curl https://raw.githubusercontent.com/…/setup.sh | bash. Piping an unsigned remote script to shell is high risk because it executes code fetched at runtime; the instruction does not pin a release or provide integrity checks.
Credentials
The SKILL.md says 'At least one LLM API key (Anthropic, OpenAI, DeepSeek, or Qwen)' and the feature list implies Telegram tokens and likely other credentials. The declared manifest lists no required env vars or primary credential — a clear inconsistency. Asking for multiple LLM provider keys and a Telegram bot token (implied) is plausible for the described features but should be declared explicitly and minimized.
Persistence & Privilege
The skill describes persistent behavior (git-managed Ring 2 repo, web dashboard, Telegram bot) and self-reproduction/self-evolution. While always:false, the runtime instructions create long-lived services and modify a local git repo, which gives significant persistence and privilege on the host. The skill also instructs executing remote setup scripts that may install persistent components.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install protea-self-evolving-life-agent - After installation, invoke the skill by name or use
/protea-self-evolving-life-agent - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of protea-Self-evolving life agent.
- Introduces a three-ring architecture: Sentinel (supervisor), Intelligence (LLM-driven evolution), and Evolvable Code (self-restructuring program).
- Supports multiple LLM providers: Anthropic, OpenAI, DeepSeek, and Qwen.
- Features code self-evolution with automatic mutation, rollback, and fitness-based survivor selection.
- Implements fitness scoring, gene pool inheritance, tiered memory management, and skill crystallization.
- Includes a Telegram bot for interaction and a local web dashboard for monitoring.
- Ships with comprehensive test coverage (1098 tests).
Metadata
Frequently Asked Questions
What is Protea Self Evolving Life Agent?
Self-evolving artificial life agent. Three-ring architecture: Ring 0 (Sentinel) supervises, Ring 1 (Intelligence) drives LLM-powered evolution, Ring 2 (Evolv... It is an AI Agent Skill for Claude Code / OpenClaw, with 443 downloads so far.
How do I install Protea Self Evolving Life Agent?
Run "/install protea-self-evolving-life-agent" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Protea Self Evolving Life Agent free?
Yes, Protea Self Evolving Life Agent is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Protea Self Evolving Life Agent support?
Protea Self Evolving Life Agent is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Protea Self Evolving Life Agent?
It is built and maintained by EdisonChenAI (@edisonchenai); the current version is v1.0.0.
More Skills