AI Skill Hub 推荐使用:adk-go-bedrock 是一款优质的MCP工具。AI 综合评分 7.1 分,在同类工具中表现稳健。如果你正在寻找可靠的MCP工具解决方案,这是一个值得深入了解的选择。
adk-go-bedrock 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
adk-go-bedrock 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
# 方式一:通过 Claude Code CLI 一键安装
claude skill install https://github.com/craigh33/adk-go-bedrock
# 方式二:手动配置 claude_desktop_config.json
{
"mcpServers": {
"adk-go-bedrock": {
"command": "npx",
"args": ["-y", "adk-go-bedrock"]
}
}
}
# 配置文件位置
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
# 安装后在 Claude 对话中直接使用 # 示例: 用户: 请帮我用 adk-go-bedrock 执行以下任务... Claude: [自动调用 adk-go-bedrock MCP 工具处理请求] # 查看可用工具列表 # 在 Claude 中输入:"列出所有可用的 MCP 工具"
// claude_desktop_config.json 配置示例
{
"mcpServers": {
"adk-go-bedrock": {
"command": "npx",
"args": ["-y", "adk-go-bedrock"],
"env": {
// "API_KEY": "your-api-key-here"
}
}
}
}
// 保存后重启 Claude Desktop 生效
<p align="center"> <img src="docs/images/readme-header.jpg" alt="adk-go-bedrock banner showing Agent Development Kit connected to Amazon Bedrock" width="100%" /> </p>
google.golang.org/adk)AWS_REGION, or region in ~/.aws/config for your profilebedrock:InvokeModel for Conversebedrock:InvokeModelWithResponseStream for ConverseStream (when ADK uses SSE streaming)go get github.com/craigh33/adk-go-bedrock
Replace the module path with your fork or published path if you rename the module in go.mod.
ctx := context.Background()
llm, err := bedrock.New(ctx, "us.anthropic.claude-3-5-sonnet-20241022-v2:0", &bedrock.Options{
Region: os.Getenv("AWS_REGION"),
})
if err != nil {
log.Fatal(err)
}
agent, err := llmagent.New(llmagent.Config{
Name: "assistant",
Model: llm,
Instruction: "You are helpful.",
})
// Wire agent into runner.New(...) as usual.
bedrock.New accepts a model ID or inference profile ARN as documented by AWS. LLMRequest.Model can override the model ID at runtime (e.g. from ADK callbacks).
The bedrock/mappers package holds genai ↔ Bedrock conversions (requests, responses, tools, usage). Import it if you need the same mappings outside the default bedrock package. It also exports MIMETypeFromExtension for inferring MIME types from a filename when building genai parts. The Bedrock Runtime API abstraction used by converse.go is exported from bedrock (RuntimeAPI, StreamReader, and NewRuntimeAPI).
These runnable programs show how to wire adk-go-bedrock into ADK agents: chat runners, tools, streaming, multimodal and document flows, guardrails, and provider-specific options. Each example has its own README.md and Makefile:
examples/bedrock-a2a: A2A remote-agent example backed by Bedrock.examples/bedrock-chat: runner-based chat example.examples/bedrock-mcp: MCP support via ADK's mcptoolset with an in-memory MCP server (MCP support).examples/bedrock-tool-calling: tool-calling agent example with function declarations.examples/bedrock-function-tool: ADK functiontool with typed args/results and llmagent + runner (contrasts with hand-written genai.Tool in bedrock-tool-calling).examples/bedrock-image-gen: ADK runner with the imagegenerator tool—Nova Canvas image generation via Bedrock InvokeModel and artifact storage.examples/bedrock-video-gen: ADK runner with the videogenerator tool—Nova Reel text-to-video via Bedrock async invoke, S3 output, and optional MP4 download into artifacts.examples/bedrock-nova-grounding: Nova Web Grounding via tools/novagrounding; prints answers and bedrock_citations metadata (see Nova Web Grounding).examples/bedrock-stream: direct streaming example using GenerateContent(..., true).examples/bedrock-tool-variants: function declaration support plus early detection of non-function ADK tool variants that Bedrock does not currently support.examples/bedrock-multimodal: comprehensive image analysis, document processing, tool calling with rich media, and vision-based reasoning.examples/bedrock-prompt-cache: ModelOption / Bedrock prompt caching for fewer repeated tokens (see AWS prompt caching docs).examples/bedrock-document: CLI to debug document uploads (-dry-run mapper check, optional -combined / -stream).examples/bedrock-guardrails: safety assessments, content filtering, and guardrail metadata handling.examples/bedrock-system-instruction: system instructions for role definition, output formatting, and behavioral control.examples/bedrock-web-ui: ADK local web UI launcher.All examples load AWS configuration with config.LoadDefaultConfig and require BEDROCK_MODEL_ID plus region configuration (AWS_REGION or profile region).
export BEDROCK_MODEL_ID=us.anthropic.claude-3-5-sonnet-20241022-v2:0
export AWS_REGION=us-east-1 # optional if your profile already defines region
make -C examples/bedrock-chat run
Run streaming example:
make -C examples/bedrock-stream run
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。
✅ 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。
总体来看,adk-go-bedrock 是一款质量良好的MCP工具,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。
| 原始名称 | adk-go-bedrock |
| 原始描述 | 开源MCP工具:Bedrock Converse Model Provider and Utilities for adk-go。⭐12 · Go |
| Topics | mcpadkadk-goai-agentsai-toolsaws-bedrockgo |
| GitHub | https://github.com/craigh33/adk-go-bedrock |
| License | Apache-2.0 |
| 语言 | Go |
收录时间:2026-05-22 · 更新时间:2026-05-22 · License:Apache-2.0 · 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