Autogen Skill

Microsoft AutoGen - 多智能体协同框架，用于构建复杂游戏设计工作流

Security Scan

The skill is internally consistent with its description: it's an instruction-only wrapper for Microsoft AutoGen and does not request extra credentials, installs, or unrelated access.

Purpose & Capability

Name/description (Microsoft AutoGen multi-agent framework for game design) match the SKILL.md usage examples which import autogen and create agents. Required capabilities (none declared) are proportional to the stated purpose.

Instruction Scope

SKILL.md gives concrete Python usage (import autogen, create AssistantAgent/UserProxyAgent, initiate_chat). It references a filesystem path (/workspace/skills/gamedev-tools/autogen/) and suggests pip package 'pyautogen' — these are reasonable but are assumptions (preinstalled path/pip name) the runtime must meet. The example config uses an LLM model ('gpt-4'), which implies model access but the skill does not declare credentials (this is common when platform supplies model access).

Install Mechanism

No install spec and no code files — lowest-risk instruction-only skill. The README claims the package is preinstalled; nothing will be written to disk by the skill itself.

Credentials

The skill requests no environment variables, secrets, or config paths. That aligns with an instruction-only wrapper. Note: using an LLM (gpt-4) requires model access from the platform but no unexpected credentials are requested by the skill.

Persistence & Privilege

always is false and default autonomous invocation is allowed (normal). The skill does not request persistent system presence or to modify other skills.

Assessment

This skill is an instruction-only wrapper for Microsoft AutoGen and appears coherent. Before installing/use: verify the referenced package/path actually exist in your runtime (/workspace/skills/gamedev-tools/autogen/ or 'pyautogen'), confirm you are comfortable giving AutoGen agents access to the platform model (the example uses 'gpt-4'), and if you need to install the package yourself, prefer the official Microsoft repo releases. If you need tighter control, test in an isolated environment since multi-agent frameworks can run autonomous interactions that trigger external API calls or code execution depending on configuration.

Microsoft AutoGen 多智能体框架的 OpenClaw 技能封装。

安装

已预装在 /workspace/skills/gamedev-tools/autogen/

使用

import autogen

# 创建助手
assistant = autogen.AssistantAgent(
    name="game_designer",
    llm_config={"model": "gpt-4"}
)

# 创建用户代理
user_proxy = autogen.UserProxyAgent(
    name="user",
    human_input_mode="NEVER"
)

# 开始对话
user_proxy.initiate_chat(
    assistant,
    message="设计一个RPG游戏的第一章剧情"
)

路径

• 源码: /workspace/skills/gamedev-tools/autogen/
• Python包: 通过 pip install pyautogen 安装