← 返回 Skills 市场
404
总下载
2
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install arena-system
功能描述
Adversarial self-improvement for AI agents. Reduces hallucinations through Agent vs Anti-Agent debate loops.
使用说明 (SKILL.md)
Arena System
Adversarial self-improvement framework for AI agents.
What it does
Give one agent two personas:
- Agent - Does the work, writes reports
- Anti-Agent - Questions everything, writes counter-reports
They take turns critiquing each other until you stop them.
Why use it
AI agents are overconfident. They hallucinate. Arena forces them to question their own outputs by arguing with themselves.
Setup
./setup.sh ~/my-arena
Creates:
my-arena/
├── state.json
├── prompts/agent.md
├── prompts/anti-agent.md
└── outputs/
Usage
Add to HEARTBEAT.md:
- Read
state.json→ whose turn? - Run that persona
- Write to
outputs/{role}/iteration_N.md - Switch turns, save state
Config
state.json:
{
"current_turn": "agent",
"iteration": 0,
"topic": "my-project",
"active": true,
"max_iterations": 10
}
Results
Prevents premature deployments, catches bugs, forces proper validation before going live.
安全使用建议
This skill appears coherent and limited to a local on-disk debate loop. Before installing: (1) inspect and, if needed, edit the prompts so they don't contain or ask for secrets; (2) run setup.sh with a directory you control (e.g., ./setup.sh ~/safe/path) rather than accepting defaults; (3) be aware outputs/state files may contain sensitive or proprietary text you feed into the loop — store them appropriately; (4) note the minor path/documentation inconsistencies (examples use ~/my-arena vs. heartbeat snippet's ~/clawd/arena) and correct them to match your setup; (5) if you want upstream provenance, verify the GitHub repository URL in package.json before trusting published claims.
功能分析
Type: OpenClaw Skill
Name: arena-system
Version: 1.0.0
The OpenClaw AgentSkills bundle 'arena-system' is designed for adversarial self-improvement of AI agents. The `setup.sh` script performs standard directory and file creation within a user-specified or default `~/clawd/arena` directory. The instructions in `SKILL.md` and the generated `heartbeat-snippet.md` guide the AI agent to read and write files (`state.json`, `prompts/*.md`, `outputs/*/*.md`) exclusively within this designated arena directory to manage the debate loop. There is no evidence of malicious intent, such as data exfiltration, unauthorized remote execution, persistence mechanisms, or prompt injection attempts designed to subvert the agent for harmful purposes. All operations are confined and directly support the stated purpose of the skill.
能力评估
Purpose & Capability
Name/description (Agent vs Anti-Agent debate loop) match the SKILL.md and the included files. The setup script and prompts create the state, prompts, and outputs the skill needs; there are no unrelated credentials, binaries, or external services requested.
Instruction Scope
SKILL.md instructs the agent to read/write local state.json and outputs and to alternate persona prompts — this is exactly the advertised behavior. Note: the loop stores whatever you feed it (reports may contain sensitive data), and the heartbeat snippet hardcodes a '~/clawd/arena/state.json' path which is a minor inconsistency with examples that use custom directories (e.g., ~/my-arena).
Install Mechanism
No install spec; the only executable provided is setup.sh which creates local folders and prompt files. No downloads, no external package installs, and nothing is written outside the user-specified directory.
Credentials
No environment variables, secrets, or external credentials are requested or referenced. The skill's file I/O (state.json, prompts, outputs) is proportionate to its purpose.
Persistence & Privilege
always is false and the skill does not request elevated or persistent platform privileges. It only writes files under the chosen directory and does not modify other skills or global agent configuration.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install arena-system - 安装完成后，直接呼叫该 Skill 的名称或使用
/arena-system触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
Initial release: adversarial self-improvement for AI agents
元数据
常见问题
Arena System 是什么？
Adversarial self-improvement for AI agents. Reduces hallucinations through Agent vs Anti-Agent debate loops. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 404 次。
如何安装 Arena System？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install arena-system」即可一键安装，无需额外配置。
Arena System 是免费的吗？
是的，Arena System 完全免费（开源免费），可自由下载、安装和使用。
Arena System 支持哪些平台？
Arena System 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Arena System？
由 Zedit42（@zedit42）开发并维护，当前版本 v1.0.0。
推荐 Skills