What makes Nvidia Agent Fleet notable?

976 downloads. Trust evidence available

How should Nvidia Agent Fleet be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

Nvidia Agent Fleet

NVIDIA Agent Fleet — 19个模型Agent + 智能调度引擎 + 并行协作 Skill: Nvidia Agent Fleet Owner: clementgu Summary: NVIDIA Agent Fleet — 19个模型Agent + 智能调度引擎 + 并行协作 Tags: latest:1.1.2 Version history: v1.1.2 | 2026-06-30T00:49:00.833Z | auto nvidia-agent-fleet v1.1.2 - Updated agent registry (agents/registry.json) for improvements or corrections. - Removed the skill-card.md file. - No public-facing feature or documentation changes. v1.1.1 | 2026-05-04T02:43:10.782Z | auto nvidia-a

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

976

Updated

Oct 11, 2026

Version

1.1.2

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 976 downloads reported by the source. Last updated 10/11/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 11, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 11, 2026
Adoption signal: 976 downloadsadoption · observed Oct 11, 2026
Latest release: 1.1.2release · observed Jun 30, 2026
Handshake status: UNKNOWNsecurity

What changed

Newest first. Only changes that point at a source are listed.

Release 1.1.2release
nvidia-agent-fleet v1.1.2 - Updated agent registry (agents/registry.json) for improvements or corrections. - Removed the skill-card.md file. - No public-facing feature or documentation changes.
Observed Jun 30, 2026 · source

Install and run

Setup complexity: low.

clawhub skill install s17fsz3rx9wrnw593kpx5w87jn85c6qc:nvidia-agent-fleet

Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.
Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-clementgu-nvidia-agent-fleet/snapshot"

Run-check

$0.02 USD

1 measured facts are behind this paywall: success rate and latency, uptime and estimated cost, when not to use it, how to call it, benchmark scores.

Agents pay $0.02 in USDC. A card payment is $0.50, the smallest a card allows.

Documentation

CLAWHUB

56,549 characters of source documentation, loaded on request.

Extracted files

4 files captured from the source.

SKILL.md

---
name: nvidia-agent-fleet
slug: nvidia-agent-fleet
version: 1.1.0
description: NVIDIA Agent Fleet — 19个模型Agent + 智能调度引擎 + 并行协作
author: openclaw
tags: [nvidia, multi-agent, dispatcher, llm, orchestration, parallel]
---

# 🚀 NVIDIA Agent Fleet

多 Agent 调度系统。每个 NVIDIA 模型拥有专属 Agent 身份，调度器自动分析任务类型并分配最佳 Agent 执行。

## 架构

```
任务输入 → 调度引擎 → 任务分类 → Agent 匹配 → 执行
                                            ├── DeepSeek V3.2 🧠
                                            ├── Qwen Coder 32B 💻
                                            ├── Kimi K2 🇨🇳
                                            ├── Llama 4 Maverick 🦙
                                            └── ... 19 个 Agent
```

## Agent 阵容 (19个)

| Agent | 模型 | 专长 |
|-------|------|------|
| 🧠 deepseek-v3-2 | DeepSeek V3.2 | 推理、逻辑、数学 |
| 🇨🇳 kimi-k2 | Kimi K2 | 中文、长文本、知识 |
| 🤔 kimi-k2-thinking | Kimi K2 Thinking | 深度思考、推理 |
| 🏆 mistral-large-3 | Mistral Large 3 (675B) | 通用智能、多语言 |
| ⚡ mistral-small | Mistral Small 4 | 快速响应 |
| 💻 qwen-coder-32b | Qwen2.5 Coder 32B | 代码生成、算法 |
| 🦾 qwen3-coder-480b | Qwen3 Coder 480B | 架构设计、审查 |
| 🔧 deepseek-coder | DeepSeek Coder 6.7B | 轻量代码、SQL |
| ⚙️ codestral | Codestral 22B | 代码补全、填空 |
| 🦙 llama-3-3-70b | Llama 3.3 70B | 通用对话 |
| 🦙🆕 llama-4-maverick | Llama 4 Maverick | 创意、分析 |
| 🔬 gemma-3-27b | Gemma 3 27B | 学术、科学 |
| 🏎️ phi-4-mini | Phi-4 Mini | 极速响应 |
| 💨 gemma-3-4b | Gemma 3 4B | 超轻量 |
| 🌏 yi-large | Yi Large | 中文优化 |
| 🏛️ glm-5-1 | GLM 5.1 | 中文问答 |
| 🐉 qwen-3-5-397b | Qwen 3.5 397B | 超大中文 |
| 👁️ llama-vision-90b | Llama Vision 90B | 视觉理解 |
| 📐 nv-embed | NV-EmbedQA | 文本嵌入 |

## 使用

### CLI

```bash
# 自动调度最佳 Agent
fleet "用Python写一个快速排序"

# 分析任务会匹配哪些 Agent
fleet analyze "解释一下量子纠缠"

# 指定 Agent 执行
fleet agent qwen-coder-32b "实现二分查找"

# 多 Agent 协同
fleet multi "如何看待人工智能的未来？"

# 列出所有 Agent
fleet list
```

### ⚡ 并行执行（v1.1.0 新增）

```bash
# 多 Agent 并行协同（同时调不同模型，互不等待）
fleet --multi --parallel "对比茅台和五粮液的财务状况"

# 指定 Agent 并行
fleet --multi --parallel --agent deepseek-v3-2,kimi-k2 "分析这个创业方案"

# 仍支持原有串行模式
fleet --multi "分析市场趋势"
```

**并行 vs 串行 速度对比**
```
串行: 任务A→任务B→任务C→任务D = 4个加起来
并行: 同时发4个请求 = 最慢那个决定总时间
加速比: x2~4 (取决于模型)
```

### 自动超时保护

每个模型有独立超时时间，快的先出结果，慢的到点自动跳过：

| 模型 | 超时 | 说明 |
|------|------|------|
| Llama 4 Maverick | 15s | 写作/报告类，极快 |
| Qwen Coder 32B | 20s | 代码/数据处理 |
| Kimi K2 | 25s | 中文分析 |
| DeepSeek V3.2 | 35s | 深度推理（最慢） |
| 其他 | 30s | 通用 |

### API Key 自动发现

支持多级检索，无需手动设置环境变量：
```
1. $NVIDIA_API_KEY (环境变量)
2. ~/.zshrc (常用 shell 配置)
3. openclaw.json (OpenClaw 配置)
```

### Python SDK

```python
from dispatcher.fleet import dispatch, select_agents, multi_dispatch

# 自动匹配
result = dispatch("实现一个Web服务器")
print(result["content"])

# 多Agent协同（串行）
results = multi_dispatch("哲学问题：意识是什么？")

# 多Agent协同（并行执行）🚀
results = multi_dispatch("分析市场", parallel=True)
for r in results:
    print(f"{r['emoji']} {r['name']}: {r['content'][:100]}...")
```

## v1.1.0 变更

### ✨ 新增
- **并行执行模式** (`--parallel` / `-p`): 多 Agent 同时调用，互不等待
- **逐模型超时保护**: 每个模型有独立超时，慢的不拖

_meta.json

{
  "ownerId": "kn7ey9xh63pt0m136bgmx0s67n85dkt5",
  "slug": "nvidia-agent-fleet",
  "version": "1.1.2",
  "publishedAt": 1782780540833
}

skill-card.md

## Description:

Routes user tasks to a registry of NVIDIA API-backed model agents, with automatic task classification, selected-agent execution, and optional multi-agent parallel responses.

This skill is ready for commercial/non-commercial use.

## Publisher:

[clementgu](https://clawhub.ai/user/clementgu)

### License/Terms of Use:

MIT-0

## Use Case:

Developers and agent users use this skill to route prompts to task-specific NVIDIA API model agents for coding, reasoning, Chinese-language tasks, writing, research, quick summaries, and multi-agent comparisons.

### Deployment Geography for Use:

Global

## Known Risks and Mitigations:

Risk: Automatic credential discovery can read local credential files and run shell configuration while looking for NVIDIA_API_KEY.

Mitigation: Set NVIDIA_API_KEY explicitly and avoid installing or running the skill in environments where shell startup files have side effects.

Risk: Prompt content is transmitted to NVIDIA's API during model calls.

Mitigation: Do not submit secrets, regulated data, or private business data unless the deployment has approved handling for that data.

## Reference(s):

- [ClawHub skill page](https://clawhub.ai/clementgu/skills/nvidia-agent-fleet)
- [Publisher profile](https://clawhub.ai/user/clementgu)
- [NVIDIA API endpoint](https://integrate.api.nvidia.com/v1)

## Skill Output:

**Output Type(s):** [text, markdown, code, shell commands, configuration, guidance]

**Output Format:** [CLI text output and Python dictionaries containing model responses, usage data, selected agents, and timing metadata]

**Output Parameters:** [1D]

**Other Properties Related to Output:** [Outputs depend on selected model agents, NVIDIA_API_KEY availability, configured max token limits, and per-model request timeouts.]

## Skill Version(s):

1.1.2 (source: server release metadata; artifact frontmatter says 1.1.0)

## Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

agents/registry.json

{
  "deepseek-v3-2": {
    "model": "deepseek-ai/deepseek-v3.2",
    "name": "DeepSeek V3.2",
    "emoji": "🧠",
    "specialty": "深度推理、逻辑分析、数学问题",
    "system_prompt": "你是一个深度推理专家。擅长逻辑分析、数学推导和复杂问题分解。给出逐步推理过程后再给出结论。",
    "strength": 0.95,
    "speed": "中等",
    "max_tokens": 4096
  },
  "kimi-k2": {
    "model": "moonshotai/kimi-k2-instruct",
    "name": "Kimi K2",
    "emoji": "🇨🇳",
    "specialty": "中文理解、长文本分析、知识问答",
    "system_prompt": "你是一个中文AI助手Kimi，擅长中文理解和知识问答。用流畅的中文回答，注重准确性和深度。",
    "strength": 0.92,
    "speed": "中等",
    "max_tokens": 8192
  },
  "kimi-k2-thinking": {
    "model": "moonshotai/kimi-k2-thinking",
    "name": "Kimi K2 Thinking",
    "emoji": "🤔",
    "specialty": "深度思考、复杂推理、多步骤决策",
    "system_prompt": "你是一个深度思考型AI。面对复杂问题，先展示完整的思考链条，再进行回答。注重推理的严谨性。",
    "strength": 0.93,
    "speed": "慢",
    "max_tokens": 4096
  },
  "mistral-large-3": {
    "model": "mistralai/mistral-large-3-675b-instruct-2512",
    "name": "Mistral Large 3",
    "emoji": "🏆",
    "specialty": "超大模型、多语言、通用智能",
    "system_prompt": "你是Mistral Large 3，一个拥有675B参数的巨型模型。擅长多语言任务、复杂推理和知识密集型回答。",
    "strength": 0.97,
    "speed": "慢",
    "max_tokens": 4096
  },
  "mistral-small": {
    "model": "mistralai/mistral-small-4-119b-2603",
    "name": "Mistral Small 4",
    "emoji": "⚡",
    "specialty": "快速响应、通用任务、性价比",
    "system_prompt": "你是一个高效的通用AI助手。在保持回答质量的同时注重响应速度。",
    "strength": 0.85,
    "speed": "快",
    "max_tokens": 2048
  },
  "qwen-coder-32b": {
    "model": "qwen/qwen2.5-coder-32b-instruct",
    "name": "Qwen Coder 32B",
    "emoji": "💻",
    "specialty": "代码生成、算法实现、Debug",
    "system_prompt": "你是一个专业的代码助手。擅长各种编程语言的代码生成、调试、重构和算法实现。给出完整可运行的代码。",
    "strength": 0.94,
    "speed": "快",
    "max_tokens": 4096
  },
  "qwen3-coder-480b": {
    "model": "qwen/qwen3-coder-480b-a35b-instruct",
    "name": "Qwen3 Coder 480B",
    "emoji": "🦾",
    "specialty": "大规模代码、架构设计、代码审查",
    "system_prompt": "你是一个顶级代码架构师。擅长大规模系统设计、代码审查和复杂重构。提供架构级别的建议。",
    "strength": 0.96,
    "speed": "慢",
    "max_tokens": 4096
  },
  "deepseek-coder": {
    "model": "deepseek-ai/deepseek-coder-6.7b-instruct",
    "name": "DeepSeek Coder",
    "emoji": "🔧",
    "specialty": "轻量代码、快速修复、SQL查询",
    "system_prompt": "你是一个轻量级代码助手。擅长快速代码生成、bug修复、SQL查询和脚本编写。",
    "strength": 0.82,
    "speed": "极快",
    "max_tokens": 2048
  },
  "codestral": {
    "model": "mistralai/codestral-22b-instruct-v0.1",
    "name": "Codestral 22B",
    "emoji": "⚙️",
    "specialty": "代码补全、填空、多语言编程",
    "system_prompt": "你是Codestral，一个专业的代码补全和生成助手。擅长代码填空、函数补全和多语言编程任务。",
    "strength": 0.88,
    "speed": "快",
    "max_tokens": 4096
  },
  "stock-analyst": {
    "model": "qwen/qwen3.5-122b-a10b",
    "name": "股票分析助手",
    "emoji": "📊",
    "specialty": "A股分析、六维框架、股票诊断",
    "system_prompt": "你是专业的A股股票分析助手。请使用六维分析框架（基本面、技术面、消息面、资金面、心理面、宏观面）对股票进行全面分析。给出具体的指标数据和操作建议。注意中国A股市场的特点。",
    "strength": 0.88,
    "speed": "快",
    "max_tokens": 2048
  },
  "llama-3-3-70b": {
    "model": "meta/llama-3.3

Github ReposUpdated 3d agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.

MCPOPENCLAW

Github ReposUpdated 7mo agoRank 70

CopilotKit

The Frontend for Agents & Generative UI. React + Angular