io.github.j0hanz/filesystem-context

数据与存储

by j0hanz

🔍 Read-only MCP server for secure filesystem exploration, searching, and analysis

什么是 io.github.j0hanz/filesystem-context

🔍 Read-only MCP server for secure filesystem exploration, searching, and analysis

README

Filesystem MCP Server

License npm version Build GitHub stars

Install in VS Code Install in VS Code Insiders Install in Visual Studio Install in Cursor

Overview

Filesystem-MCP is a Model Context Protocol server that lets AI assistants read and write files within explicitly allowed directories. Sensitive file patterns (.env, *.pem, *id_rsa*) are blocked by default. It exposes filesystem tools, resources, and prompts over stdio or Streamable HTTP transport.

AspectDetails
StatusActive (see npm badge for the current version)
LanguageTypeScript (strict)
RuntimeNode.js >= 24
Packagenpm
LicenseMIT

Features

FeatureDescription
Path guardingEvery path is validated against allowed roots; .env, *.pem, *id_rsa* and similar patterns are denied
Filesystem toolsNavigate, inspect, read, and write across all major file operations
Batch operationsMost tools accept path, paths[], or files[] for parallel execution
Dual transportstdio by default; --port enables Streamable HTTP
File subscriptionsResource subscriptions push change notifications when watched files update
Regex safetyRE2 in all search tools: linear-time matching, so no pattern can ReDoS the server

Built with

Node.js TypeScript Docker

LayerTechnology
ProtocolMCP SDK v2 (@modelcontextprotocol/server)
RuntimeNode.js >= 24 · TypeScript 6 · ESM
Transportstdio (default) · Streamable HTTP (--port)
RegexRE2 (re2-wasm) — linear time, no lookahead/lookbehind/backreferences
ContainerDocker alpine · multi-stage build · non-root user

Table of Contents

Quick start

[!NOTE] Requires Node.js ≥ 24.

Prerequisites

RequirementVersion / Notes
Node.js≥ 24
npmBundled with Node.js
DockerOptional — for container use

Install via npx

bash
npx -y @j0hanz/filesystem-mcp /path/to/allowed/dir

Or install globally:

bash
npm install -g @j0hanz/filesystem-mcp
filesystem-mcp /path/to/allowed/dir

Install via Docker

bash
docker run -i --rm \
  -v /path/to/project:/workspace:ro \
  ghcr.io/j0hanz/filesystem-mcp:latest \
  --read-only /workspace

Configure in VS Code

Add to .vscode/mcp.json:

json
{
  "servers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@j0hanz/filesystem-mcp@latest", "/path/to/project"]
    }
  }
}

Or install via CLI:

sh
code --add-mcp '{"name":"filesystem","command":"npx","args":["-y","@j0hanz/filesystem-mcp@latest","/path/to/project"]}'

Configure in Visual Studio

Add to .vs\mcp.json in your solution directory, or %USERPROFILE%\.mcp.json for a global configuration:

json
{
  "servers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@j0hanz/filesystem-mcp@latest", "/path/to/project"]
    }
  }
}

Configure in Claude Desktop

Add to your claude_desktop_config.json:

json
{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@j0hanz/filesystem-mcp@latest", "/path/to/project"]
    }
  }
}

Install in Cursor

Add to .cursor/mcp.json in your project root (project-scoped), or ~/.cursor/mcp.json for a global configuration:

json
{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@j0hanz/filesystem-mcp@latest", "/path/to/project"]
    }
  }
}

Docker configuration

VS Code (.vscode/mcp.json) and Visual Studio (.vs\mcp.json):

json
{
  "servers": {
    "filesystem": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-v",
        "/path/to/project:/workspace",
        "ghcr.io/j0hanz/filesystem-mcp:latest",
        "/workspace"
      ]
    }
  }
}

Claude Desktop (claude_desktop_config.json) and Cursor (mcp.json):

json
{
  "mcpServers": {
    "filesystem": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-v",
        "/path/to/project:/workspace",
        "ghcr.io/j0hanz/filesystem-mcp:latest",
        "/workspace"
      ]
    }
  }
}

[!NOTE] For least privilege, use both controls: :ro makes the container mount read-only at the operating-system boundary, while the server's --read-only flag removes mutating tools (create, edit, move, delete, patch, replace_text) from tools/list.

Usage

Tools

All tools are scoped to the configured roots. Call list_roots first to discover what is allowed.

Navigate

ToolDescription
list_rootsList allowed workspace roots. Call this first — all other tools scope to these.
listList directory contents. Returns entries (dirs-first, alphabetical) and an ASCII tree.
find_filesFind files by glob pattern (e.g. **/*.ts). Returns matching files with metadata.

Inspect

ToolDescription
statGet file/directory metadata: size, modified time, permissions, MIME type, token estimate.
search_textSearch file contents for text (grep-like). Returns matching lines with context.
diffCompare two files and return a unified diff with added/removed line counts.

Read

ToolDescription
readRead a text file. Supports head/tail and line ranges. Accepts paths[] for batches.

Write

ToolDescription
createCreate one or more files, overwriting existing content and creating parent directories as needed.
editApply sequential literal string replacements to one or more files (max 5 per call).
moveMove, rename, or copy (copy: true) one or more files/directories to explicit destinations.
deletePermanently delete one or more files or directories. This action is irreversible.
replace_textBulk search-and-replace across files matching a glob pattern.
patchApply a single-file unified diff and write the result.

Resources

URIDescription
internal://instructionsServer navigation guide — tools overview, constraints, and error recovery.
filesystem-mcp://file/{+path}Read a workspace file. Subscribe to receive push notifications on change.
filesystem-mcp://result/{id}Ephemeral cached tool output. Expires after ~60 seconds, eviction, or server restart.

Prompts

PromptDescription
get-helpReturn usage instructions, optionally filtered to a specific section.

Project structure

text
filesystem-mcp/
├── __tests__/        Test suites
├── src/
│   ├── core/         Path guarding, filesystem abstraction, concurrency, observability
│   ├── tools/        Tool definitions and registration
│   ├── index.ts      Process entrypoint and transport selection
│   ├── server.ts     Server factory and registrar composition
│   ├── transport/    stdio and Streamable HTTP transport setup
│   ├── prompts.ts    Prompt definitions and registration
│   └── resources.ts  Resource definitions and registration
└── Dockerfile        Multi-stage alpine build, non-root user

Runtime composition flows from src/index.ts to src/transport.ts, then to src/server.ts, the registrars, and finally src/core/. Each registrar owns the narrow dependency contract it consumes.

PathPurpose
src/core/path.tsPathGuard — validates every path against allowed roots
src/core/fs.tsGuardedFileSystem — guarded filesystem facade
src/tools/define.tsTool registration and execution framework
src/tools/batch.tsBatch helpers (runOverPaths, isTotalFailure)
src/server.tsBuilds shared dependencies and invokes the three registrars
src/transport.tsOwns stdio and Streamable HTTP setup around the server factory

Configuration

The server starts with allowed directories from explicit startup configuration:

  1. Positional directories passed to filesystem-mcp.
  2. Environment variable FS_ALLOWED_DIRS (separated by : on POSIX or ; on Windows).
  3. Current working directory when --allow-cwd is enabled.

Legacy MCP connections may additionally seed roots through the deprecated roots/list flow. Modern 2026-07-28 connections do not automatically send workspace roots. They can add access after startup by calling a tool with a concrete path and approving the elicitation-backed grant. list_roots reports the roots already configured or accepted; it cannot discover an unknown workspace by itself.

Recommended global recipes

VS Code / Cursor / Claude Code (primary recipe)

Configure the project directory explicitly:

Add to your global or project-scoped configuration:

json
{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@j0hanz/filesystem-mcp@latest", "/path/to/project"]
    }
  }
}

Claude Desktop (fallback recipe via environment variable)

Claude Desktop and similar clients don't support the MCP Roots protocol. Use the FS_ALLOWED_DIRS environment variable to configure allowed folders.

Add to your claude_desktop_config.json:

json
{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@j0hanz/filesystem-mcp@latest"],
      "env": {
        "FS_ALLOWED_DIRS": "/path/to/project1:/path/to/project2"
      }
    }
  }
}

(On Windows, separate directories with a semicolon ; instead of a colon :).

Advanced / per-project positional arguments

You can also restrict access to specific directories by passing positional arguments directly:

bash
# Start with explicit positional paths
filesystem-mcp /path/to/project1 /path/to/project2

Configuration reference

CLI flags

FlagDefaultPurpose
[dirs...]One or more allowed root directories (positional)
--allow-cwdfalseAlso allow the current working directory as a root
--walk-cwdfalseWalk up from CWD to find a project root; implies --allow-cwd
--allow-missing-rootsfalseStart even if configured allowed directories do not exist
--port <n>Enable Streamable HTTP transport on the given port (env: FS_PORT)
--http-host <host>HTTP server bind address (env: FS_HTTP_HOST)
--api-key <key>Require this API key on HTTP requests (env: FS_API_KEY)
--read-onlyfalseDisable write tools: create, edit, delete, move, patch, replace_text
--safefalseAlias for --read-only
--deny <pattern>Block paths matching this pattern; repeatable
--allow-sensitivefalseAllow access to sensitive system paths (env: FS_ALLOW_SENSITIVE)
--root-boundary <path>Require all allowed roots to fall under this path (env: FS_ROOT_BOUNDARY)
--max-file-size <bytes>Maximum file size for reads in bytes (env: FS_MAX_FILE_SIZE)
--log-level <level>infoRFC 5424 log level, debug through emergency (env: FS_LOG_LEVEL)
--print-configfalsePrint the active configuration and exit (use --json for machine-readable output)
--jsonfalseOutput --print-config as JSON

Environment variables

All boolean variables accept true or 1 to enable and false, 0, or unset to disable; any other value logs a warning and reads as disabled. Flags take precedence when both are set.

VariablePurpose
FS_ALLOWED_DIRSColon-separated (POSIX) or semicolon-separated (Windows) list of directories to allow.
FS_ROOT_BOUNDARYPath prefix all allowed roots must fall under (mirrors --root-boundary).
FS_ALLOW_CWD_WALKWalk up from CWD to find a project root (mirrors --walk-cwd).
FS_ALLOW_MISSING_ROOTSStart even if configured directories do not exist (mirrors --allow-missing-roots).
FS_ALLOW_SENSITIVEAllow access to sensitive system paths (mirrors --allow-sensitive).
FS_DENYLISTComma-separated list of paths or patterns to block (mirrors --deny).
FS_MAX_FILE_SIZEMaximum file size for reads in bytes (mirrors --max-file-size).
FS_LOG_LEVELRFC 5424 log level: debug, info, notice, warn/warning, error, critical, alert, or emergency (mirrors --log-level).
FS_PORTStart the Streamable HTTP transport on this port; unset = stdio (mirrors --port).
FS_HTTP_HOSTHTTP server bind address (mirrors --http-host).
FS_API_KEYAPI key required on HTTP requests (mirrors --api-key).
FS_TRUST_PROXYExpress trust proxy setting: hop count or expression. Unset = do not trust X-Forwarded-*.
FS_ALLOWED_HOSTSComma-separated Host header values to accept (HTTP transport).
FS_ALLOWED_ORIGINSComma-separated origin hostnames for CORS.
FS_ALLOW_UNRESTRICTED_HOSTSBind a wildcard host with no Host validation (accepts the risk).
FS_PUBLIC_URLResource identifier URL for RFC 9728 discovery.
FS_RATE_LIMIT_RPMPer-client-IP requests/minute (default 120 with API-key authentication, 6,000 for keyless loopback; range 1–100000).
FS_MAX_REQUEST_BYTESMax HTTP request body bytes (default 4194304, 1024–268435456).
FS_KEEPALIVE_TIMEOUT_MSHTTP keep-alive timeout in ms; set above any fronting proxy's idle timeout (default 5000, 1000–600000).
FS_MAX_WATCHERSMax concurrent file watchers (default 256, 1–4096).
FS_MAX_INLINE_MATCHESDeprecated and ignored; maxResults sets the search_text page size. Logs a warning when set; removed in the next major.
FS_MAX_READ_MANY_BYTESMax total bytes across a batched read (default 524288, 10240–104857600).
FS_SEARCH_TIMEOUT_MSSearch timeout in ms (default 5000, 100–60000).
NO_COLORAny value disables ANSI color output.
FS_REQUEST_STATE_KEYHMAC key sealing input_required requestState across retry rounds. Optional (random per boot if unset); set it, at >=32 bytes UTF-8, to keep in-flight rounds alive across a restart.

Examples

bash
# Allow current working directory
filesystem-mcp --allow-cwd

# HTTP transport on port 3000
filesystem-mcp --port 3000

Scripts

ModeCommandDescription
Full checknpm run checkRun build, type check, lint, format, knip, and tests
Auto-fix + checknpm run fixAuto-fix formatting/linting and run the full check
Static onlynpm run check:staticRun static analysis without tests
Tests onlynpm testRun tests; accepts native node --test options

Security

[!IMPORTANT] Report vulnerabilities privately via GitHub Security Advisories. Do not open public issues for security reports.

TopicDetail
Path traversalEvery path is resolved and validated against allowed roots before any operation
Sensitive files.env, *.pem, *id_rsa*, and similar patterns are denied by default
Regex safetyRE2 cannot backtrack, so a hostile pattern cannot hang the server (ReDoS)
ContainerRuns as non-root mcp user; bind mounts control what is exposed

Contributing

  1. Fork the repository.
  2. Create a feature branch: git checkout -b feat/your-feature.
  3. Commit your changes with a clear message.
  4. Run npm run check to confirm tests, types, lint, formatting, and knip all pass.
  5. Open a pull request.

Contributors

License

Released under the MIT License. See LICENSE for details.

常见问题

io.github.j0hanz/filesystem-context 是什么?

🔍 Read-only MCP server for secure filesystem exploration, searching, and analysis

相关 Skills

技术栈评估

by alirezarezvani

Universal
热门

对比框架、数据库和云服务,结合 5 年 TCO、安全风险、生态活力与迁移复杂度做量化评估,适合技术选型、栈升级和替换路线决策。

帮你系统比较技术栈优劣,不只看功能,还把TCO、安全性和生态健康度一起量化,选型和迁移决策更稳。

数据与存储
未扫描26.0k

资深数据科学家

by alirezarezvani

Universal
热门

覆盖实验设计、特征工程、预测建模、因果推断与模型评估,适合用 Python/R/SQL 做 A/B 测试、时序分析和生产级 ML 落地,支撑数据驱动决策。

从 A/B 测试、因果分析到预测建模一条龙搞定,既有硬核统计方法也懂业务沟通,特别适合把数据结论真正落地。

数据与存储
未扫描26.0k

资深架构师

by alirezarezvani

Universal
热门

适合系统设计评审、ADR记录和扩展性规划,分析依赖与耦合,权衡单体或微服务、数据库与技术栈选型,并输出Mermaid、PlantUML、ASCII架构图。

搞系统设计、技术选型和扩展规划时,用它能更快理清架构决策与依赖关系,还能直接产出 Mermaid/PlantUML 图,方案讨论效率很高。

数据与存储
未扫描26.0k

相关 MCP Server

by Anthropic

热门

PostgreSQL 是让 Claude 直接查询和管理你的数据库的 MCP 服务器。

这个服务器解决了开发者需要手动编写 SQL 查询的痛点,特别适合数据分析师或后端开发者快速探索数据库结构。不过,由于是参考实现,生产环境使用前务必评估安全风险,别指望它能处理复杂事务。

数据与存储
89.7k

SQLite 数据库

编辑精选

by Anthropic

热门

SQLite 是让 AI 直接查询本地数据库进行数据分析的 MCP 服务器。

这个服务器解决了 AI 无法直接访问 SQLite 数据库的问题,适合需要快速分析本地数据集的开发者。不过,作为参考实现,它可能缺乏生产级的安全特性,建议在受控环境中使用。

数据与存储
89.2k

by Firecrawl

热门

Firecrawl 是让 AI 直接抓取网页并提取结构化数据的 MCP 服务器。

它解决了手动写爬虫的麻烦,让 Claude 能直接访问动态网页内容。最适合需要实时数据的研究者或开发者,比如监控竞品价格或抓取新闻。但要注意,它依赖第三方 API,可能涉及隐私和成本问题。

数据与存储
7.1k

评论