经 AI Skill Hub 精选评估,人工智能红队引擎 获评「强烈推荐」。这款Agent工作流在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 8.0 分,适合有一定技术背景的用户使用。
人工智能红队引擎 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
人工智能红队引擎 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:pip 安装(推荐)
pip install humanbound
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install humanbound
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/humanbound/humanbound
cd humanbound
pip install -e .
# 验证安装
python -c "import humanbound; print('安装成功')"
# 命令行使用
humanbound --help
# 基本用法
humanbound input_file -o output_file
# Python 代码中调用
import humanbound
# 示例
result = humanbound.process("input")
print(result)
# humanbound 配置文件示例(config.yml) app: name: "humanbound" debug: false log_level: "INFO" # 运行时指定配置文件 humanbound --config config.yml # 或通过环境变量配置 export HUMANBOUND_API_KEY="your-key" export HUMANBOUND_OUTPUT_DIR="./output"
<p align="center"> <img src="https://raw.githubusercontent.com/humanbound/humanbound/main/assets/logo-dark.svg" alt="Humanbound" width="280"/> </p>
<p align="center"> Open-source adversarial testing engine, SDK, and CLI for AI agents. <br/> Attack your agent the way real users and attackers will: live endpoints, multi-turn conversations, tool abuse. Then turn every failure into a firewall rule. <br/> Runs locally or against the Humanbound Platform. No login required to start. </p>
<p align="center"> <a href="#quick-start">Quick Start</a> · <a href="#from-test-results-to-guardrails">Test-to-Guardrail Loop</a> · <a href="#python-sdk">SDK</a> · <a href="https://docs.humanbound.ai/">Documentation</a> · <a href="#contributing">Contributing</a> </p>
<p align="center"> <a href="https://pypi.org/project/humanbound/"><img src="https://img.shields.io/pypi/v/humanbound?style=flat-square&color=FD9506" alt="PyPI version"/></a> <a href="https://pypi.org/project/humanbound/"><img src="https://img.shields.io/pypi/pyversions/humanbound?style=flat-square&color=FD9506" alt="Python versions"/></a> <a href="https://pypi.org/project/humanbound/"><img src="https://img.shields.io/pypi/dm/humanbound?style=flat-square&color=FD9506" alt="Downloads"/></a> <a href="https://github.com/humanbound/humanbound/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/humanbound/humanbound/ci.yml?style=flat-square&color=FD9506" alt="CI"/></a> <a href="https://github.com/humanbound/humanbound/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-Apache--2.0-FD9506?style=flat-square" alt="License"/></a> <a href="https://discord.gg/WgTMpmSFtN"><img src="https://img.shields.io/badge/discord-community-FD9506?style=flat-square" alt="Discord"/></a> <a href="https://docs.humanbound.ai/"><img src="https://img.shields.io/badge/docs-humanbound.ai-FD9506?style=flat-square" alt="Docs"/></a> </p>
---
📖 Full documentation lives at docs.humanbound.ai — this README covers the essentials; the docs have the depth.
- Clean name: humanbound is the PyPI install. The old humanbound-cli package has been yanked from PyPI; install humanbound directly. - Public SDK namespace alongside the CLI — use the CLI or drive the engine from Python. - Firewall integration: pip install humanbound[firewall] pulls humanbound-firewall alongside the CLI.
See the changelog for full release notes.
pip install humanbound # CLI + SDK, core deps
pip install humanbound[engine] # + OpenAI / Anthropic / Gemini providers
pip install humanbound[firewall] # + humanbound-firewall runtime
pip install humanbound[engine,firewall] # everything
```bash
export HB_PROVIDER=openai export HB_API_KEY=sk-...
import json
import time
from humanbound import LocalRunner, TestConfig
config = TestConfig(
endpoint=json.load(open("bot-config.json")), # same config file the CLI uses
scope_path="scope.yaml",
)
runner = LocalRunner()
experiment_id = runner.start(config)
while runner.get_status(experiment_id).status not in ("Finished", "Failed", "Terminated"):
time.sleep(5)
posture = runner.get_posture(experiment_id)
print(f"Security posture: {posture.overall_score} ({posture.grade})")
for insight in runner.get_result(experiment_id).insights:
print(insight["severity"], "—", insight["category"])
The CLI and SDK share the same implementation, so they cannot drift. Authoring custom orchestrators and wiring EngineCallbacks is covered in the docs.
humanbound 是一个开源的 AI 代理红队引擎、SDK 和 CLI。它可以在本地运行或与其他系统进行交互。
humanbound 2.0 版本带来了以下新功能:使用 humanbound 作为 PyPI 安装包,公开的 SDK 命名空间,CLI 和 Python SDK 共享同一实现,防火墙集成等。
要安装 humanbound,可以使用以下命令: pip install humanbound(CLI 和 SDK,核心依赖) pip install humanbound[engine](+ OpenAI / Anthropic / Gemini 提供者) pip install humanbound[firewall](+ humanbound-firewall 运行时) pip install humanbound[engine,firewall](所有内容)
CLI 使用示例 See docs.humanbound.ai for the full example
要配置 LLM 提供者,可以使用以下环境变量: export HB_PROVIDER=openai export HB_API_KEY=sk-...
Python SDK from humanbound import Bot, LocalRunner, OwaspAgentic, TestingLevel, EngineCallbacks
Compose your own test pipeline bot = Bot(endpoint="https://my-agent/chat", api_key="...") class Callbacks(EngineCallbacks): def on_finding(self, insight): ... def on_progress(self, pct): ... runner = LocalRunner()
高质量的开源AI工作流项目,适合AI安全和红队测试
该工具使用 NOASSERTION 协议,商用场景请仔细阅读协议条款,必要时咨询法律意见。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
📄 NOASSERTION — 请查阅原始协议条款了解具体使用限制。
AI Skill Hub 点评:人工智能红队引擎 的核心功能完整,质量优秀。对于自动化工程师和运维人员来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。
| 原始名称 | humanbound |
| 原始描述 | 开源AI工作流:Open-source AI agent red-team engine, SDK, and CLI. Run offline or against the H。⭐27 · Python |
| Topics | AI红队测试AI安全Python |
| GitHub | https://github.com/humanbound/humanbound |
| License | NOASSERTION |
| 语言 | Python |
收录时间:2026-06-09 · 更新时间:2026-06-11 · License:NOASSERTION · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端