# Awesome Mcp For Coding Agents

> A curated list of Model Context Protocol servers for coding agents — Claude Code, Cursor, Cline, Aider, Windsurf.

- **Type:** MCP server
- **Install:** `agentstack add mcp-swarmclawai-awesome-mcp-for-coding-agents`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [swarmclawai](https://agentstack.voostack.com/s/swarmclawai)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [swarmclawai](https://github.com/swarmclawai)
- **Source:** https://github.com/swarmclawai/awesome-mcp-for-coding-agents

## Install

```sh
agentstack add mcp-swarmclawai-awesome-mcp-for-coding-agents
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# awesome-mcp-for-coding-agents

> A curated list of Model Context Protocol (MCP) servers that are genuinely useful to **coding agents** — Claude Code, Cursor, Cline, Aider, Windsurf.

[](https://awesome.re)
[](./LICENSE)
[](https://modelcontextprotocol.io/)

## Why this list

There are hundreds of MCP servers out there. Most of them aren't made for coding — they're Slack bots, CRMs, calendar integrations, and smart-home bridges. When you're trying to configure Claude Code or Cursor to actually help you *ship code*, a generic catalog wastes hours.

This list is different:

- **Coding focus only.** Every entry helps an agent read code, write code, run code, debug code, or understand a system the developer is building against.
- **Maintained only.** Each entry had a commit within ~90 days of its last review. Stale repos are removed on each pass.
- **Install-ready.** Every entry ships with the one-line command to run it via stdio.
- **Opinionated.** ⭐ marks an editor's pick — the one to install first in a category.

If an entry rots or gets archived, open a PR and we'll swap it out. See [CONTRIBUTING.md](./CONTRIBUTING.md) for the rules.

## Contents

1. [Filesystem & Code Search](#1-filesystem--code-search)
2. [Git & GitHub/GitLab](#2-git--githubgitlab)
3. [Browser & Web](#3-browser--web)
4. [Databases & Data](#4-databases--data)
5. [Testing, Errors & Observability](#5-testing-errors--observability)
6. [Docs & Knowledge](#6-docs--knowledge)
7. [Memory & Context](#7-memory--context)
8. [Infra & Dev Tools](#8-infra--dev-tools)

[How to install an MCP server (Claude Code · Cursor · Cline · Aider · Windsurf)](#how-to-install) · [Contributing](./CONTRIBUTING.md)

---

## 1. Filesystem & Code Search

- ⭐ **[server-filesystem](https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem)** — Official reference filesystem server with scoped directory access, read/write, search, and edit tools. `npx -y @modelcontextprotocol/server-filesystem /path/to/allowed/dir`
- **[mark3labs/mcp-filesystem-server](https://github.com/mark3labs/mcp-filesystem-server)** — Go implementation of the filesystem server with the same tool surface; single static binary, useful when you don't want Node. `go install github.com/mark3labs/mcp-filesystem-server@latest`
- ⭐ **[oraios/serena](https://github.com/oraios/serena)** — LSP-powered semantic code toolkit: symbol search, find-references, safe renames across a repo — the closest thing to giving an agent a real IDE. `uvx --from git+https://github.com/oraios/serena serena start-mcp-server`
- **[isaacphi/mcp-language-server](https://github.com/isaacphi/mcp-language-server)** — Generic LSP bridge (gopls, pyright, rust-analyzer, tsserver, etc.) exposing definition/references/diagnostics/rename to the agent. `go install github.com/isaacphi/mcp-language-server@latest`
- **[wonderwhy-er/DesktopCommanderMCP](https://github.com/wonderwhy-er/DesktopCommanderMCP)** — Filesystem + terminal + diff-based code editing in one server; very popular with Claude Desktop users. `npx -y @wonderwhy-er/desktop-commander@latest setup`
- **[tumf/mcp-text-editor](https://github.com/tumf/mcp-text-editor)** — Line-oriented text editor server optimized for LLM token efficiency (patch-style edits with hash-based optimistic locking). `uvx mcp-text-editor`
- **[mrjoshuak/godoc-mcp](https://github.com/mrjoshuak/godoc-mcp)** — Go-specific: feeds the agent `go doc` output for any package/symbol instead of making it read source files. `go install github.com/mrjoshuak/godoc-mcp@latest`

## 2. Git & GitHub/GitLab

- ⭐ **[github/github-mcp-server](https://github.com/github/github-mcp-server)** — Official GitHub MCP: issues, PRs, reviews, files, actions, code search across repos. `docker run -i --rm -e GITHUB_PERSONAL_ACCESS_TOKEN ghcr.io/github/github-mcp-server`
- **[server-git (reference)](https://github.com/modelcontextprotocol/servers/tree/main/src/git)** — Official Python git server: status, diff, log, show, commit, branch ops on a local repo. `uvx mcp-server-git --repository /path/to/repo`
- **[cyanheads/git-mcp-server](https://github.com/cyanheads/git-mcp-server)** — Node/TS git server with broader command coverage (worktrees, stash, rebase, cherry-pick) than the reference impl. `npx -y @cyanheads/git-mcp-server`
- ⭐ **[idosal/git-mcp](https://github.com/idosal/git-mcp)** — Turns any public GitHub repo into an MCP endpoint so the agent can Q&A/search that codebase as docs without cloning. `npx -y mcp-remote https://gitmcp.io//`
- **[zereight/gitlab-mcp](https://github.com/zereight/gitlab-mcp)** — GitLab MCP with MR/issue/pipeline/wiki tools; the de-facto GitLab option since the reference impl was archived. `npx -y @zereight/mcp-gitlab`
- **[circleci-public/mcp-server-circleci](https://github.com/circleci-public/mcp-server-circleci)** — Official CircleCI server: fetch failing builds, logs, and flaky tests so the agent can debug CI without leaving the editor. `npx -y @circleci/mcp-server-circleci`
- **[sooperset/mcp-atlassian](https://github.com/sooperset/mcp-atlassian)** — Jira + Confluence; useful when the agent needs ticket context or has to update an issue after a fix. `uvx mcp-atlassian`

## 3. Browser & Web

- ⭐ **[microsoft/playwright-mcp](https://github.com/microsoft/playwright-mcp)** — Official Microsoft Playwright server; structured accessibility-tree snapshots rather than pixels — the standard choice for web automation. `npx -y @playwright/mcp@latest`
- ⭐ **[ChromeDevTools/chrome-devtools-mcp](https://github.com/ChromeDevTools/chrome-devtools-mcp)** — Official Google server exposing Chrome DevTools (console, network, performance, DOM) to agents; ideal for debugging a running web app. `npx -y chrome-devtools-mcp@latest`
- **[browserbase/mcp-server-browserbase](https://github.com/browserbase/mcp-server-browserbase)** — Cloud headless browser sessions via Browserbase + Stagehand; good when local Chromium is awkward. `npx -y @browserbase/mcp-server-browserbase`
- **[AgentDeskAI/browser-tools-mcp](https://github.com/AgentDeskAI/browser-tools-mcp)** — Pairs with a Chrome extension to let the agent read console logs, network requests, and the DOM of your *actual* browser tab. `npx -y @agentdeskai/browser-tools-mcp@latest`
- **[hangwin/mcp-chrome](https://github.com/hangwin/mcp-chrome)** — Chrome-extension–based MCP that reuses your existing logged-in browser (cookies, sessions) — handy for scraping authed sites. See repo for extension + bridge install.
- **[server-fetch (reference)](https://github.com/modelcontextprotocol/servers/tree/main/src/fetch)** — Minimal URL fetch with HTML→markdown conversion. `uvx mcp-server-fetch`
- ⭐ **[firecrawl/firecrawl-mcp-server](https://github.com/firecrawl/firecrawl-mcp-server)** — Firecrawl-powered scraping, crawling, and extraction with markdown output; strong for docs/web reading. `npx -y firecrawl-mcp`
- **[tavily-ai/tavily-mcp](https://github.com/tavily-ai/tavily-mcp)** — Tavily web search + extract API; optimized-for-LLM results. `npx -y tavily-mcp`
- **[exa-labs/exa-mcp-server](https://github.com/exa-labs/exa-mcp-server)** — Exa neural web search with content retrieval and research tools. `npx -y exa-mcp-server`
- **[nickclyde/duckduckgo-mcp-server](https://github.com/nickclyde/duckduckgo-mcp-server)** — Free, keyless web search via DuckDuckGo. `uvx duckduckgo-mcp-server`
- **[ihor-sokoliuk/mcp-searxng](https://github.com/ihor-sokoliuk/mcp-searxng)** — Privacy-respecting SearXNG meta-search; point at any SearXNG instance. `npx -y mcp-searxng`

## 4. Databases & Data

- ⭐ **[bytebase/dbhub](https://github.com/bytebase/dbhub)** — Universal DB gateway speaking Postgres, MySQL, SQL Server, SQLite, MariaDB, Oracle via one MCP — the "one server for every DB" pick. `docker run --rm -i bytebase/dbhub --transport stdio --dsn `
- **[crystaldba/postgres-mcp](https://github.com/crystaldba/postgres-mcp)** — "Postgres Pro": schema introspection, read-only query, EXPLAIN, index advice, performance diagnostics — purpose-built for coding/debugging. `pipx install postgres-mcp`
- **[supabase-community/supabase-mcp](https://github.com/supabase-community/supabase-mcp)** — Official Supabase: branch/migrate/query and generate TypeScript types for Supabase projects. `npx -y @supabase/mcp-server-supabase@latest --access-token `
- **[neondatabase/mcp-server-neon](https://github.com/neondatabase/mcp-server-neon)** — Neon Postgres: create branches, run migrations, query — great for schema work without leaving the agent. `npx -y @neondatabase/mcp-server-neon start `
- **[mongodb-js/mongodb-mcp-server](https://github.com/mongodb-js/mongodb-mcp-server)** — Official MongoDB: query, aggregate, manage indexes, Atlas cluster ops. `npx -y mongodb-mcp-server --connectionString `
- **[redis/mcp-redis](https://github.com/redis/mcp-redis)** — Official Redis: CRUD across strings/hashes/streams/JSON, plus Redis Stack (vector search, time series). `uvx --from redis-mcp-server@latest redis-mcp-server --url redis://localhost:6379`
- **[ClickHouse/mcp-clickhouse](https://github.com/ClickHouse/mcp-clickhouse)** — Official ClickHouse: schema browsing and read-only SELECT — great for exploring analytics data mid-debug. `uv run --with mcp-clickhouse --python 3.13 mcp-clickhouse`
- **[motherduckdb/mcp-server-motherduck](https://github.com/motherduckdb/mcp-server-motherduck)** — DuckDB / MotherDuck; lets the agent SQL-query local files (Parquet/CSV) or cloud DuckDB. `uvx mcp-server-motherduck --db-path md:`
- **[Snowflake-Labs/mcp](https://github.com/Snowflake-Labs/mcp)** — Official Snowflake: Cortex Search/Analyst, object management, query execution. `uvx snowflake-labs-mcp --service-config-file config.yaml`
- **[elastic/mcp-server-elasticsearch](https://github.com/elastic/mcp-server-elasticsearch)** — Official Elastic: search indices, run DSL/ES|QL queries — useful when the agent debugs log-heavy apps. `docker run -i --rm -e ES_URL -e ES_API_KEY docker.elastic.co/mcp/elasticsearch stdio`
- **[dbt-labs/dbt-mcp](https://github.com/dbt-labs/dbt-mcp)** — Official dbt: explore models, run compile/build, ask the semantic layer. `uvx dbt-mcp`

## 5. Testing, Errors & Observability

- ⭐ **[getsentry/sentry-mcp](https://github.com/getsentry/sentry-mcp)** — Official Sentry: pull error events, issue details, stack traces, and Seer autofix suggestions into the agent's context. `npx -y @sentry/mcp-server@latest`
- **[grafana/mcp-grafana](https://github.com/grafana/mcp-grafana)** — Official Grafana: search dashboards, query Prometheus/Loki, list alerts, fetch incidents. `uvx mcp-grafana`
- **[pab1it0/prometheus-mcp-server](https://github.com/pab1it0/prometheus-mcp-server)** — Prometheus MCP: PromQL, metric discovery, target inspection — debugging with real metrics on tap. `docker run -i --rm -e PROMETHEUS_URL ghcr.io/pab1it0/prometheus-mcp-server:latest`
- **[circleci-public/mcp-server-circleci](https://github.com/circleci-public/mcp-server-circleci)** — Official CircleCI: failing-build analysis, flaky-test detection, pipeline config validation. `npx -y @circleci/mcp-server-circleci`
- **[tumf/mcp-shell-server](https://github.com/tumf/mcp-shell-server)** — Allow-listed shell runner; the minimal way to let the agent invoke `pytest`, `vitest`, `go test`, etc. `uvx mcp-shell-server`
- **[g0t4/mcp-server-commands](https://github.com/g0t4/mcp-server-commands)** — Simpler "run any command" server, with per-call confirmation hooks. `npx mcp-server-commands`

## 6. Docs & Knowledge

- ⭐ **[upstash/context7](https://github.com/upstash/context7)** — Up-to-date, version-pinned library docs injected on demand ("use context7") — fixes hallucinated APIs. `npx -y @upstash/context7-mcp`
- ⭐ **[langchain-ai/mcpdoc](https://github.com/langchain-ai/mcpdoc)** — Serve any site's `llms.txt` / `llms-full.txt` as MCP docs; ideal for framework maintainers and polyglot projects. `uvx --from mcpdoc mcpdoc --urls "LangGraph:https://langchain-ai.github.io/langgraph/llms.txt"`
- **[arabold/docs-mcp-server](https://github.com/arabold/docs-mcp-server)** — Self-hosted docs indexer that crawls library sites and exposes semantic search; good offline/private alternative to Context7. `docker run --rm -p 6280:6280 -v docs-mcp-data:/data ghcr.io/arabold/docs-mcp-server:latest`
- **[regenrek/deepwiki-mcp](https://github.com/regenrek/deepwiki-mcp)** — Query DeepWiki's generated docs for any GitHub repo. `npx -y mcp-deepwiki@latest`
- **[apollographql/apollo-mcp-server](https://github.com/apollographql/apollo-mcp-server)** — Turn a GraphQL schema into MCP tools so the agent understands your API surface when writing queries. See repo for `apollo-mcp-server` install.
- **[mrjoshuak/godoc-mcp](https://github.com/mrjoshuak/godoc-mcp)** — Go package documentation lookup (also listed in §1).
- **[idosal/git-mcp](https://github.com/idosal/git-mcp)** — Doubles as "docs for any public repo" via code + README + wiki search. `npx -y mcp-remote https://gitmcp.io//`

## 7. Memory & Context

- **[server-memory (reference)](https://github.com/modelcontextprotocol/servers/tree/main/src/memory)** — Official knowledge-graph memory: persistent entities/relations/observations across sessions. `npx -y @modelcontextprotocol/server-memory`
- ⭐ **[basicmachines-co/basic-memory](https://github.com/basicmachines-co/basic-memory)** — Local-first markdown knowledge base the agent writes and reads; human-editable notes, no opaque store. `uvx basic-memory mcp`
- **[shaneholloman/mcp-knowledge-graph](https://github.com/shaneholloman/mcp-knowledge-graph)** — Drop-in replacement for `server-memory` with customizable on-disk graph path — handy per-project memory. `npx -y @shaneholloman/mcp-knowledge-graph --memory-path ./memory.jsonl`
- ⭐ **[qdrant/mcp-server-qdrant](https://github.com/qdrant/mcp-server-qdrant)** — Official Qdrant: store and semantically recall code snippets / past solutions as a vector memory. `uvx mcp-server-qdrant`
- **[pinecone-io/pinecone-mcp](https://github.com/pinecone-io/pinecone-mcp)** — Official Pinecone: manage indexes and do vector search. `npx -y @pinecone-database/mcp`
- **[neo4j-contrib/mcp-neo4j](https://github.com/neo4j-contrib/mcp-neo4j)** — Official Neo4j MCPs (Cypher, memory, aura ops) — graph memory the agent can traverse. `uvx mcp-neo4j-cypher@latest --db-url bolt://localhost:7687 --username neo4j --password `
- **[getzep/graphiti](https://github.com/getzep/graphiti)** — Temporal knowledge-graph engine; ships an MCP server for bi-temporal agent memory. `pip install graphiti-core` then see repo `mcp_server/`.
- **[supermemoryai/supermemory-mcp](https://github.com/supermemoryai/supermemory-mcp)** — Hosted "memory layer" that an agent can write to / read from across sessions and clients. See repo for hosted URL install.

## 8. Infra & Dev Tools

- ⭐ **[awslabs/mcp](https://github.com/awslabs/mcp)** — The AWS mono-repo of official MCP servers (CLI, CDK, Terraform, Cost Explorer, docs, Bedrock, ECS, Lambda, etc.). `uvx awslabs.core-mcp-server@latest`
- **[cloudflare/mcp-server-cloudflare](https://github.com/cloudflare/mcp-server-cloudflare)** — Official Cloudflare MCP(s) for Workers, KV, R2, D1, Pages, Durable Objects. `npx -y mcp-remote https://bindings.mcp.cloudflare.com/sse`
- **[hashicorp/terraform-mcp-server](https://github.com/hashicorp/terraform-mcp-server)** — Official HashiCorp: provider discovery, module analysis, Registry lookup for IaC work. `docker run -i --rm hashicorp/terraform-mcp-server`
- ⭐ **[Flux159/mcp-server-kubernetes](https://github.com/Flux159/mcp-server-kubernetes)** — Popular TypeScript Kubernetes MCP covering pods, deployments, services, logs, port-forward. `npx -y mcp-server-kubernetes`
- **[manusa/kubernetes-mcp-server](https://github.com/manusa/kubernetes-mcp-server)** — Alternative K8s MCP (also supports OpenShift) with CRUD over any resource + specialized tools. `npx -y kubernetes-mcp-server@latest`
- **[docker/mcp-gateway](https://github.com/docker/mcp-gateway)** — Docker's official gateway that containerizes and multiplexe

…

## Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [swarmclawai](https://github.com/swarmclawai)
- **Source:** [swarmclawai/awesome-mcp-for-coding-agents](https://github.com/swarmclawai/awesome-mcp-for-coding-agents)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** yes
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/mcp-swarmclawai-awesome-mcp-for-coding-agents
- Seller: https://agentstack.voostack.com/s/swarmclawai
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.