⚙️
Agent工作流
Gym AI工作流评估框架
基于 Python · 无代码搭建完整 AI 自动化流程
英文名：Gym
8.2AI 综合评分
AI评估工作流智能体基准测试环境模拟
✦ AI Skill Hub 推荐
AI Skill Hub 强烈推荐：Gym AI工作流评估框架 是一款优质的Agent工作流。已获得 1.0k 颗 GitHub Star，AI 综合评分 8.2 分，在同类工具中表现稳健。如果你正在寻找可靠的Agent工作流解决方案，这是一个值得深入了解的选择。
📚 深度解析
Gym AI工作流评估框架 是一套完整的 AI Agent 自动化工作流方案。随着 AI 能力的不断提升，基于 Agent 的自动化工作流正在成为提升个人和团队效率的核心方式。区别于传统的 RPA 自动化（模拟鼠标键盘操作），AI Agent 工作流通过理解任务意图、动态规划执行路径，能够处理更复杂的非结构化任务。
Gym AI工作流评估框架 工作流的设计遵循"最小配置，最大复用"原则：核心逻辑已经封装好，用户只需配置自己的 API Key 和业务参数即可快速上手。工作流内置错误处理和重试机制，在网络波动或 API 限速等情况下仍能稳定运行，适合作为生产环境的自动化基础设施。
在实际部署时，建议先在测试环境中运行 3-5 次，验证各个环节的输出结果符合预期，再部署到生产环境。AI Skill Hub 评分 8.2 分，是同类 Agent 工作流中的精选推荐。
Gym AI工作流评估框架 工作流的设计遵循"最小配置，最大复用"原则：核心逻辑已经封装好，用户只需配置自己的 API Key 和业务参数即可快速上手。工作流内置错误处理和重试机制，在网络波动或 API 限速等情况下仍能稳定运行，适合作为生产环境的自动化基础设施。
在实际部署时，建议先在测试环境中运行 3-5 次，验证各个环节的输出结果符合预期，再部署到生产环境。AI Skill Hub 评分 8.2 分，是同类 Agent 工作流中的精选推荐。
📋 工具概览
Gym AI工作流评估框架 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。
GitHub Stars
⭐ 1.0k
开发语言
Python
支持平台
Windows / macOS / Linux
维护状态
正常维护，社区驱动
开源协议
Apache-2.0
AI 综合评分
8.2 分
工具类型
Agent工作流
Forks
208
📖 中文文档
以下内容由 AI Skill Hub 根据项目信息自动整理，如需查看完整原始文档请访问底部「原始来源」。
Gym AI工作流评估框架 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。
📌 核心特色
- 可视化 Agent 工作流编排，无需编写复杂代码
- 支持多步骤自动化任务链，实现全流程无人值守
- 与外部 API、数据库和第三方服务无缝集成
- 内置错误处理与自动重试机制，保障稳定运行
- 提供可复用的自动化模板，快速在同类场景部署
🎯 主要使用场景
- 自动化日常重复性工作，将精力集中于创造性任务
- 构建数据采集 → 处理 → 输出的完整自动化管线
- 实现跨平台、跨系统的数据流转和业务协同
以下安装命令基于项目开发语言和类型自动生成，实际以官方 README 为准。
安装命令
# 方式一：pip 安装（推荐）
pip install gym
# 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 gym
# 方式三：从源码安装（获取最新功能）
git clone https://github.com/NVIDIA-NeMo/Gym
cd Gym
pip install -e .
# 验证安装
python -c "import gym; print('安装成功')"
📋 安装步骤说明
- 访问 GitHub 仓库获取工作流文件
- 在对应平台（Dify / Flowise / Make 等）中找到「导入工作流」功能
- 上传工作流文件
- 按照提示配置必要的环境变量和 API Key
- 运行测试确认流程正常后投入使用
以下用法示例由 AI Skill Hub 整理，涵盖最常见的使用场景。
常用命令 / 代码示例
# 命令行使用
gym --help
# 基本用法
gym input_file -o output_file
# Python 代码中调用
import gym
# 示例
result = gym.process("input")
print(result)
以下配置示例基于典型使用场景生成，具体参数请参照官方文档调整。
配置示例
# gym 配置文件示例（config.yml） app: name: "gym" debug: false log_level: "INFO" # 运行时指定配置文件 gym --config config.yml # 或通过环境变量配置 export GYM_API_KEY="your-key" export GYM_OUTPUT_DIR="./output"
以下内容由系统直接从 GitHub README 解析整理，保留代码块、表格与列表结构。
NeMo Gym
Requirements • Quick Start • Environment Tutorials • Available Environments • Documentation & Resources • Community & Support • Citations
NeMo Gym is a library for evaluating and improving models and agents using environments. NeMo Gym provides infrastructure to develop environments, scalably run evaluation and training, and a collection of popular benchmarks and training environments.
An environment is the complete system an agent interacts with to complete a task. It consists of a dataset (tasks to solve), an agent harness (how the model interacts with the world), a verifier (task completion scoring), and state (per-task execution context).
📋 Requirements
NeMo Gym is designed to run on standard development machines:
| Hardware Requirements | Software Requirements |
|---|---|
| **GPU**: Not required for NeMo Gym library operation<br>• GPU may be needed for specific resources servers or model inference (see individual server documentation) | **Operating System**:<br>• Linux (Ubuntu 20.04+, or equivalent)<br>• macOS (11.0+ for x86_64, 12.0+ for Apple Silicon)<br>• Windows (via WSL2) |
| **CPU**: Any modern x86_64 or ARM64 processor (e.g., Intel, AMD, Apple Silicon) | **Python**: 3.13.14 or higher |
| **RAM**: Minimum 8 GB (16 GB+ recommended for larger environments) | **Git**: For cloning the repository |
| **Storage**: Minimum 5 GB free disk space for installation and basic usage | **Internet Connection**: Required for downloading dependencies and API access |
Additional Requirements
- API Keys: OpenAI API key with available credits (for the quickstart examples)
- Other model providers supported (Azure OpenAI, self-hosted models via vLLM)
- Ray: Automatically installed as a dependency (no separate setup required)
🚀 Quick Start
Requires Python 3.13.14+ on x86_64 or ARM64 (Linux, macOS, Windows via WSL2). No GPU required. See the Getting Started docs for a more comprehensive walkthrough.
Install NeMo Gym:
Requires uv and Python 3.13.14+.
git clone git@github.com:NVIDIA-NeMo/Gym.git
cd Gym
uv venv --python 3.13.14 && source .venv/bin/activate
uv syncConfigure your model:
This quickstart uses OpenAI. NeMo Gym supports local and hosted inference — see Configure Model for vLLM, Fireworks, OpenRouter, and others.
Create env.yaml in the project root:
policy_base_url: https://api.openai.com/v1
policy_api_key: <your-openai-api-key>
policy_model_name: gpt-4.1-2025-04-14🧭 Environment Tutorials
Learn how to build custom environments through hands-on tutorials. Here are popular starting points:
| Name | Demonstrates |
|---|---|
| [Single Step](https://docs.nvidia.com/nemo/gym/main/environment-tutorials/single-step-environment) | Basic single-step tool calling |
| [Multi Step](https://docs.nvidia.com/nemo/gym/main/environment-tutorials/multi-step-environment) | Multi-step tool calling |
| [Session State](https://docs.nvidia.com/nemo/gym/main/environment-tutorials/stateful-environment) | Session state management (in-memory) |
| [Multi Reward](https://docs.nvidia.com/nemo/gym/main/build-verifiers/multi-reward-verification) | Multiple reward components for evaluation and multi-objective RL (e.g. GDPO) |
See all environment tutorials for additional patterns and advanced topics.
📦 Available Environments
Environments for training and evaluation.
Each resources server includes example data, configuration files, and tests. See each server's README for details.
The Dataset column links to publicly available datasets (e.g., on HuggingFace). A - means the train/validation data has not been publicly released yet, or that it is procedurally generated using a provided script. If no data is released yet, new data can be generated, or the environment can be used as a reference. Each server includes 5 example tasks in data/example.jsonl.
| Environment | Domain | Description | Value | Train | Validation | License | Config | Dataset |
|---|---|---|---|---|---|---|---|---|
| Aalcr | other | - | - | - | - | - | <a href='resources_servers/aalcr/configs/aalcr.yaml'>aalcr.yaml</a> | - |
| Abstention | rlhf | Train models to abstain when unsure using three-tier reward on HotPotQA with LLM judge | Improve calibration by rewarding abstention over incorrect answers | ✓ | ✓ | Creative Commons Attribution-ShareAlike 4.0 International | <a href='resources_servers/abstention/configs/abstention.yaml'>abstention.yaml</a> | - |
| Agentif | instruction_following | AgentIF instruction-following benchmark (707 agentic scenarios) scored with CSR/ISR using an LLM judge for llm/llm_conditional_check constraints and code exec for code constraints. | Improve instruction-following in agentic scenarios across unconditional, conditional, and example-driven constraint dimensions. | - | ✓ | - | <a href='resources_servers/agentif/configs/agentif.yaml'>agentif.yaml</a> | - |
| Anyswe Agent | coding | SWE-bench run by a NeMo Fabric-selected harness inside each task environment. | Evaluate software-engineering agents with a harness-neutral Fabric lifecycle. | - | - | - | <a href='responses_api_agents/anyswe_agent/configs/anyswe_nemo_fabric.yaml'>anyswe_nemo_fabric.yaml</a> | - |
| Anyswe Agent | coding | SWE-bench run by Claude Code natively inside the task container. | Eval software engineering capabilities on SWE-bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyswe_agent/configs/anyswe_claude_code.yaml'>anyswe_claude_code.yaml</a> | - |
| Anyswe Agent | coding | SWE-bench run by Hermes Agent natively inside the task container. | Eval software engineering capabilities on SWE-bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyswe_agent/configs/anyswe_hermes.yaml'>anyswe_hermes.yaml</a> | - |
| Anyswe Agent | coding | SWE-bench run by OpenClaw natively inside the task container. | Eval software engineering capabilities on SWE-bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyswe_agent/configs/anyswe_openclaw.yaml'>anyswe_openclaw.yaml</a> | - |
| Anyswe Agent | coding | SWE-bench run by OpenCode inside the task container. | Eval software engineering capabilities on SWE-bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyswe_agent/configs/anyswe_opencode.yaml'>anyswe_opencode.yaml</a> | - |
| Anyswe Agent | coding | SWE-bench run by Pi inside the task container. | Eval software engineering capabilities on SWE-bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyswe_agent/configs/anyswe_pi.yaml'>anyswe_pi.yaml</a> | - |
| Anyswe Agent | coding | SWE-bench run by the Cline CLI natively inside the task container. | Eval software engineering capabilities on SWE-bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyswe_agent/configs/anyswe_cline.yaml'>anyswe_cline.yaml</a> | - |
| Anyterminal Agent | coding | Terminal Bench run by a NeMo Fabric harness inside the task sandbox. | Evaluate terminal-task agents through a harness-neutral Fabric lifecycle. | - | - | - | <a href='responses_api_agents/anyterminal_agent/configs/anyterminal_nemo_fabric.yaml'>anyterminal_nemo_fabric.yaml</a> | - |
| Anyterminal Agent | coding | Terminal Bench run by claude-code natively inside the task container. | Evaluate terminal-task capabilities on Terminal Bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyterminal_agent/configs/anyterminal_claude_code.yaml'>anyterminal_claude_code.yaml</a> | - |
| Anyterminal Agent | coding | Terminal Bench run by OpenClaw natively inside the task container. | Evaluate terminal-task capabilities on Terminal Bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyterminal_agent/configs/anyterminal_openclaw.yaml'>anyterminal_openclaw.yaml</a> | - |
| Anyterminal Agent | coding | Terminal Bench run by Terminus-2 natively inside the task container. | Evaluate terminal-task capabilities on Terminal Bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyterminal_agent/configs/anyterminal_terminus_2.yaml'>anyterminal_terminus_2.yaml</a> | - |
| Anyterminal Agent | coding | Terminal Bench run by the Hermes agent inside the task container. | Evaluate terminal-task capabilities on Terminal Bench with any Gym agent. | - | - | - | <a href='responses_api_agents/anyterminal_agent/configs/anyterminal_hermes.yaml'>anyterminal_hermes.yaml</a> | - |
| Arc Agi | knowledge | Solve puzzles designed to test intelligence. See https://arcprize.org/arc-agi. | Improve puzzle-solving capabilities. | - | ✓ | - | <a href='resources_servers/arc_agi/configs/arc_agi.yaml'>arc_agi.yaml</a> | - |
| Arena | rlhf | LMArena proxy v2 chat evaluation benchmark | Measure general chat quality via win rate against baseline responses | - | ✓ | - | <a href='resources_servers/arena/configs/lmarena_v2.yaml'>lmarena_v2.yaml</a> | - |
| Arena | rlhf | LMArena proxy v3 chat evaluation benchmark | Measure general chat quality via win rate against baseline responses | - | ✓ | - | <a href='resources_servers/arena/configs/lmarena_v3.yaml'>lmarena_v3.yaml</a> | - |
| Arena Judge | - | - | - | - | - | <a href='resources_servers/arena_judge/configs/arena_judge.yaml'>arena_judge.yaml</a> | - | |
| Asr With Pc | other | ASR with WER scoring (standard, case-sensitive, punctuation+capitalization) | Improve transcription quality with structural detail | - | - | - | <a href='resources_servers/asr_with_pc/configs/asr_with_pc.yaml'>asr_with_pc.yaml</a> | - |
| Bbq | safety | BBQ comparative QA with Answer and Explanation Quality checks | Evidence-grounded and fairness-safe comparative reasoning | - | - | - | <a href='resources_servers/bbq/configs/bbq.yaml'>bbq.yaml</a> | - |
| Bigcodebench | coding | Verifies model-generated Python solutions against the BigCodeBench unittest suite. | Improve practical, library-rich Python coding capabilities. | - | - | - | <a href='resources_servers/bigcodebench/configs/bigcodebench.yaml'>bigcodebench.yaml</a> | - |
| Bird Sql | coding | Text-to-SQL with execution-based evaluation on BIRD dev (1534 SQLite tasks). Binary reward from unordered result-set equality. | Improve text-to-SQL capabilities on BIRD's realistic dev split using execution-based binary reward without an LLM judge. | - | - | - | <a href='resources_servers/bird_sql/configs/bird_sql.yaml'>bird_sql.yaml</a> | - |
| Blackjack | games | Blackjack. Model hits or stands. Reward +1 win, 0 draw, -1 loss/bust. | Example gymnasium-style multi-step environment | - | - | - | <a href='resources_servers/blackjack/configs/blackjack.yaml'>blackjack.yaml</a> | - |
| Browsecomp Advanced Harness | agent | Model uses search tools to satisfy a user query. | Measure agentic search capability | - | - | - | <a href='resources_servers/browsecomp_advanced_harness/configs/browsecomp_advanced_harness.yaml'>browsecomp_advanced_harness.yaml</a> | - |
| Bunsenbench Chemistry Mcq | knowledge | Public BunsenBench chemistry multiple-choice benchmark verifier | Measure chemistry MCQ reasoning with source and taxonomy breakdowns | - | - | - | <a href='resources_servers/bunsenbench_chemistry_mcq/configs/bunsenbench_chemistry_mcq.yaml'>bunsenbench_chemistry_mcq.yaml</a> | - |
| Calendar | agent | Multi-turn calendar scheduling dataset. User states events and constraints in natural language; model schedules events to satisfy all constraints. | Improve multi-turn instruction following capabilities | ✓ | ✓ | Apache 2.0 | <a href='resources_servers/calendar/configs/calendar.yaml'>calendar.yaml</a> | <a href='https://huggingface.co/datasets/nvidia/Nemotron-RL-agent-calendar_scheduling'>Nemotron-RL-agent-calendar_scheduling</a> |
| Calendar | agent | Multi-turn calendar scheduling dataset. User states events and constraints in natural language; model schedules events to satisfy all constraints. | Improve multi-turn instruction following capabilities | ✓ | ✓ | Creative Commons Attribution 4.0 International | <a href='resources_servers/calendar/configs/calendar_v2.yaml'>calendar_v2.yaml</a> | <a href='https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Calendar-v2'>Nemotron-RL-Instruction-Following-Calendar-v2</a> |
| Circle Click | other | Click on circles in images | Improve visual grounding and spatial reasoning | - | - | - | <a href='resources_servers/circle_click/configs/circle_click.yaml'>circle_click.yaml</a> | - |
| Circle Count | other | Count circles of a given color in images | Improve visual counting and color recognition | - | - | - | <a href='resources_servers/circle_count/configs/circle_count.yaml'>circle_count.yaml</a> | - |
| Citation If | instruction_following | Citation instruction-following reward checker | Train models to follow citation format instructions in search-grounded synthesis tasks | - | - | - | <a href='resources_servers/citation_if/configs/citation_if.yaml'>citation_if.yaml</a> | - |
| Code Fim | coding | Code Fill-in-the-Middle judged by HumanEval-Infilling test suite (single_line, multi_line, random_span, random_span_light) | Improve Python code-infilling capabilities (prefix + completion + suffix) | - | - | - | <a href='resources_servers/code_fim/configs/code_fim.yaml'>code_fim.yaml</a> | - |
| Code Gen | coding | Model must submit the right code to solve a problem | Improve competitive coding capabilities | ✓ | ✓ | Apache 2.0 | <a href='resources_servers/code_gen/configs/code_gen.yaml'>code_gen.yaml</a> | <a href='https://huggingface.co/datasets/nvidia/nemotron-RL-coding-competitive_coding'>nemotron-RL-coding-competitive_coding</a> |
| Competitive Coding Challenges | coding | Execution of competitive programming competition questions | Improve competitive coding capabilities on contest-style problems | - | - | - | <a href='resources_servers/competitive_coding_challenges/configs/competitive_coding_challenges.yaml'>competitive_coding_challenges.yaml</a> | - |
| Conversational Tool Use Simulation | agent | Conversational tool-use simulation environment. | - | - | - | - | <a href='resources_servers/conversational_tool_use_simulation/configs/conversational_tool_use_simulation.yaml'>conversational_tool_use_simulation.yaml</a> | - |
| Critpt | other | Research-level physics problems scored by the Artificial Analysis API | Evaluate model performance on research-level physics reasoning | - | - | - | <a href='resources_servers/critpt/configs/critpt.yaml'>critpt.yaml</a> | - |
| Equivalence Llm Judge | agent | Short bash command generation questions with LLM-as-a-judge | Improve foundational bash and IF capabilities | ✓ | ✓ | GNU General Public License v3.0 | <a href='resources_servers/equivalence_llm_judge/configs/nl2bash-equivalency.yaml'>nl2bash-equivalency.yaml</a> | - |
| Equivalence Llm Judge | knowledge | Short answer questions with LLM-as-a-judge | Improve knowledge-related benchmarks like GPQA / HLE | - | - | - | <a href='resources_servers/equivalence_llm_judge/configs/equivalence_llm_judge.yaml'>equivalence_llm_judge.yaml</a> | - |
| Equivalence Rule | knowledge | Question - Answering with rule-based reward | Improve retrieval and counting capabilities | - | - | - | <a href='resources_servers/equivalence_rule/configs/lc.yaml'>lc.yaml</a> | - |
| Ether0 | knowledge | ether0 chemistry benchmark verifiers | Evaluate chemistry knowledge and reasoning with ether0 benchmark | - | ✓ | - | <a href='resources_servers/ether0/configs/ether0.yaml'>ether0.yaml</a> | - |
| Evalplus | coding | Function-completion code judged by EvalPlus base + plus tests (HumanEval+, MBPP+) | Improve Python function-completion capabilities | - | - | - | <a href='resources_servers/evalplus/configs/evalplus.yaml'>evalplus.yaml</a> | - |
| Finance Agent V2 |
🎯 aiskill88 AI 点评
A 级
2026-06-30
成熟的AI评估框架，1k星标体现社区认可度。工作流和基准测试功能完整，适合规模化模型评估。代码质量和维护状态良好。
📚 实用指南（长尾问题）
适合谁
- 需要 Gym 解决具体问题的开发者与运营人员
最佳实践
- 先在测试环境跑通最小用例，再接入生产数据
常见错误
- API key 直接提交到 git 仓库（请用 .env 并加入 .gitignore）
- Python 依赖冲突：建议用 venv / uv 隔离环境
部署方案
- 云端托管：可放在 Vercel / Railway / Fly.io 等 PaaS 平台
⚡ 核心功能
- 可视化 Agent 工作流编排，无需编写复杂代码
- 支持多步骤自动化任务链，实现全流程无人值守
- 与外部 API、数据库和第三方服务无缝集成
- 内置错误处理与自动重试机制，保障稳定运行
- 提供可复用的自动化模板，快速在同类场景部署
👥 适合谁
- 需要 Gym 解决具体问题的开发者与运营人员
⭐ 最佳实践
- 先在测试环境跑通最小用例，再接入生产数据
⚠️ 常见错误
- API key 直接提交到 git 仓库（请用 .env 并加入 .gitignore）
- Python 依赖冲突：建议用 venv / uv 隔离环境
👥 适合人群
🎯 使用场景
- 自动化日常重复性工作，将精力集中于创造性任务
- 构建数据采集 → 处理 → 输出的完整自动化管线
- 实现跨平台、跨系统的数据流转和业务协同
⚖️ 优点与不足
✅ 优点
- +Apache-2.0 协议，可免费商用
- +大幅减少重复性人工操作
- +可视化流程，清晰直观
- +可扩展性强，支持复杂场景
⚠️ 不足
- −初始配置和调试需投入一定时间
- −强依赖外部服务的稳定性
- −复杂场景需具备一定技术基础
⚠️ 使用须知
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。
📄 License 说明
✅ Apache 2.0 — 宽松开源协议，可商用，需保留版权声明和 NOTICE 文件，含专利授权条款。
🔗 相关工具推荐
📚 相关教程推荐
📰 相关 AI 新闻
🍿 AI 圈相关吃瓜
🗺️ 相关解决方案
🧩 你可能还需要
基于当前 Skill 的能力图谱，自动补全的工具组合
❓ 常见问题 FAQ
Gym 是一款Python开发的AI辅助工具。开源AI工作流：Evaluate and improve models and agents using environments。⭐1.0k · Python 主要应用场景包括：模型性能评估、智能体训练优化、AI工作流编排。
💡 AI Skill Hub 点评
总体来看，Gym AI工作流评估框架 是一款质量优秀的Agent工作流，在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态，建议收藏备用，结合自身场景选择合适时机引入使用。
🌐 原始信息
| 原始名称 | Gym |
| 原始描述 | 开源AI工作流：Evaluate and improve models and agents using environments。⭐1.0k · Python |
| Topics | AI评估工作流智能体基准测试环境模拟 |
| GitHub | https://github.com/NVIDIA-NeMo/Gym |
| License | Apache-2.0 |
| 语言 | Python |
🔗 原始来源
🐙 GitHub 仓库 https://github.com/NVIDIA-NeMo/Gym
🌐 官方网站 https://docs.nvidia.com/nemo/gym/main/about/
收录时间：2026-06-30 · 更新时间：2026-07-03 · License：Apache-2.0 · AI Skill Hub 不对第三方内容的准确性作法律背书。