s3-mcp-connector
Enables agents to interact with any S3-compatible storage (AWS S3, LocalStack, MinIO) via MCP, providing tools like list, read, write, delete objects and buckets.
README
<div align="center">
🪣 s3-mcp-connector
Talk to Amazon S3 — or LocalStack, MinIO, any S3-compatible store — from an MCP-speaking agent.
</div>
A single static Go binary that speaks the Model Context Protocol and exposes 8 tools for working with S3: list buckets, list/read/write/delete objects, create/delete buckets. Point it at real AWS or at a local LocalStack/MinIO instance with one environment variable — same binary, same tools, zero code changes.
No Python, no uv, no runtime dependency to install — just a binary and
an .mcp.json.
✨ Why this exists
An agent that can only talk about your S3 buckets isn't that useful. This gives it hands: it can look inside a bucket, read a config file out of it, drop a report back in, or clean up a stale prefix — safely, with guardrails on size and destructive actions built in.
🧰 Tools
| Tool | What it does | Write? |
|---|---|---|
s3_list_buckets |
List buckets visible to the credentials | |
s3_list_objects |
List objects in a bucket, optional prefix, paginated | |
s3_head_object |
Object metadata (size, type, ETag) without downloading | |
s3_get_object |
Read an object's content — text or base64, size-capped | |
s3_put_object |
Write a small text/base64 object (≤ 5 MB) | ✍️ |
s3_delete_object |
Delete one object | 🗑️ destructive |
s3_create_bucket |
Create a bucket | ✍️ |
s3_delete_bucket |
Delete an empty bucket | 🗑️ destructive |
Every tool accepts an optional response_format: markdown (default,
pretty tables for a chat UI) or json (for programmatic use).
s3_get_object auto-detects text vs. binary content and caps output at
200,000 bytes by default (raise via max_bytes, hard cap
5,000,000) — it's built for reading configs, logs, and small data
files, not bulk transfer. Reach for the AWS CLI or SDK directly for large
objects.
🚀 Quickstart
Fastest path: grab a prebuilt bundle from the latest release —
download s3-mcp-connector-plugin-<version>-<os>-<arch>.zip, unzip it,
and point Cowork/Claude at the plugin/ folder inside (see step 4 of
SETUP.md). No Go toolchain required.
From source:
# 1. Build
cd go-server
go mod tidy
go build -o s3-connector-server .
cp s3-connector-server ../plugin/servers/go/
# 2. Spin up LocalStack to test against (no AWS account needed)
cd ..
docker compose up -d
# 3. Point the connector at it
export AWS_ACCESS_KEY_ID=test
export AWS_SECRET_ACCESS_KEY=test
export AWS_REGION=us-east-1
export S3_ENDPOINT_URL=http://localhost:4566
export S3_FORCE_PATH_STYLE=true
./go-server/s3-connector-server # serves MCP over stdio
Or make build && make localstack-up — see the Makefile for
every shortcut (test, vet, fmt, lint, tidy, localstack-down).
Full walkthrough — including wiring this up as a Claude/Cowork plugin and switching from LocalStack to real AWS — is in SETUP.md.
🔐 Configuration
Everything is environment variables, passed through by the plugin's
.mcp.json:
| Variable | Purpose | Default |
|---|---|---|
AWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY |
AWS credentials. Any non-empty values work against LocalStack. | — |
AWS_REGION |
Region to use. | us-east-1 |
S3_ENDPOINT_URL |
Custom endpoint. Leave unset for real AWS. | (unset) |
S3_FORCE_PATH_STYLE |
true for LocalStack/MinIO (path-style addressing). |
false |
🧪 Quality bar
This isn't a toy script — it's got the same checks you'd expect from a production Go service:
- ✅ Unit tests for every input-validation path (
go test ./...) - ✅
go vet+gofmtclean - ✅ golangci-lint (govet, staticcheck, errcheck, gosec, and more)
- ✅ govulncheck — no known vulnerabilities in the dependency graph
- ✅ CodeQL static security analysis on every push
- ✅ End-to-end verified against real LocalStack — every tool (create/delete bucket, put/get/head/list/delete object, and the 404 error path) was exercised against a live S3-compatible service, not mocks
- ✅ Dependabot keeps Go modules and Actions current
All of it runs in CI on every push and PR.
🏷️ Releases & versioning
Versions follow semver and are cut automatically by
release-please from
Conventional Commits on main:
fix: ...→ patch (v0.1.0→v0.1.1)feat: ...→ minor (v0.1.1→v0.2.0)feat!: .../BREAKING CHANGE:footer → major (v0.2.0→v1.0.0)
Every merged PR updates a standing "chore(main): release vX.Y.Z" PR with an auto-generated CHANGELOG.md. Merging that PR:
- tags the release and publishes it on GitHub
- builds and attaches zipped, ready-to-install plugin bundles for linux/darwin/windows × amd64/arm64
- regenerates
server.jsonfrom those exact assets (fresh version + SHA-256 hashes) and publishes it to the official MCP Registry viamcp-publisher, authenticated with GitHub OIDC — no stored secrets
See .github/workflows/release-please.yml
and .github/workflows/publish-mcp-registry.yml
(also runnable by hand for an existing tag via workflow_dispatch).
📁 Layout
s3-mcp-connector/
├── README.md ← you are here
├── SETUP.md ← step-by-step setup guide (LocalStack + real AWS)
├── CONTRIBUTING.md ← how to contribute
├── CODE_OF_CONDUCT.md
├── SECURITY.md ← vulnerability reporting
├── CODEOWNERS
├── LICENSE ← MIT
├── Makefile ← build / test / lint / localstack shortcuts
├── docker-compose.yml ← LocalStack, for local testing
├── .golangci.yml ← lint rules
├── release-please-config.json ← semver/changelog automation config
├── .release-please-manifest.json
├── server.json ← MCP Registry manifest (regenerated fresh per release by CI)
├── scripts/
│ └── render-server-json.sh ← rebuilds server.json from a release's zip assets
├── .github/
│ ├── workflows/
│ │ ├── ci.yml ← build, vet, test, lint, govulncheck
│ │ ├── codeql.yml ← security scanning
│ │ ├── pr-title.yml ← Conventional Commits PR title check
│ │ ├── release-please.yml ← version PRs, tagging, GitHub releases
│ │ ├── publish-mcp-registry.yml ← publishes server.json to the MCP Registry
│ │ └── rebuild-release-assets.yml ← manual re-attach fallback
│ ├── ISSUE_TEMPLATE/
│ ├── PULL_REQUEST_TEMPLATE.md
│ └── dependabot.yml
├── go-server/ ← the MCP server source
│ ├── main.go
│ ├── main_test.go
│ ├── go.mod / go.sum
│ └── README.md
└── plugin/ ← installable Cowork/Claude plugin
├── .claude-plugin/plugin.json
├── .mcp.json ← holds credentials locally — never commit real ones
└── servers/go/ ← compiled binary goes here
🤝 Contributing
PRs and issues are very welcome — see CONTRIBUTING.md for the full guide (setup, coding conventions, how to add a new tool) and the Code of Conduct.
main is protected: every change, including the maintainer's, lands via
pull request with CI green. PR titles must follow
Conventional Commits — that's what
drives the automatic versioning above.
Found a security issue? Please follow SECURITY.md instead of opening a public issue.
📄 License
MIT © Ferhat Dundar
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。