项目说明
🧠 300+ Ad Open-Source Repositories for Professional & Senior Engineers [2026]
<p align="center">
<a href=" src=" alt="Stars"></a>
<img src=" alt="Senior / Pro">
<img src=" alt="300+ repos">
<img src=" alt="Production grade">
<a href="CONTRIBUTING.md"><img src=" alt="PRs welcome"></a>
<a href="LICENSE"><img src=" alt="Apache 2.0"></a>
</p>
<p align="center"><em>No fluff, no beginner detours - just the repositories senior engineers actually reach for. 🧠⚙️<br/>Distributed systems, cloud-nati, the full LLM-engineering stack, and a serious security arsenal.</em></p>
---
> A deep, battle-tested catalog of 300+ production-grade repositories for senior software engineers, staff engineers, SREs, platform & infrastructure engineers, security engineers, ML/AI engineers, and data engineers.
>
> 🎯 Audience: Professionals & Ad. This is the "senior engineer" companion to the 280+ comprehensive list (for e) and the 150 essentials list (for beginners & students). E here assumes you already ship software - it's about depth: distributed systems, cloud-nati, obser, LLM engineering, and offensi security.
> 🗓️ Curated on 21 July 2026 from li across GitHub Trending, the CNCF landscape, Papers-we-Love, and practitioner communities. Every link is a real, canonical GitHub repository.
---
📍 Table of Contents
🏛️ System Design & Distributed Systems🗄️ Databases & Storage Engines🌊 Data Engineering & Pipelines📡 Message Queues, Streaming & E⚡ LLM Serving & Inference🧩 LLM App Frameworks & Orchestration🤖 AI Agents & Multi-Agent Systems🔎 RAG, Databases & Agent Memory🎛️ Model Fine-Tuning & Training🧪 LLM E, Obser & Guardrails📦 MLOps & Experiment Tracking⚙️ DevOps & Infrastructure-as-Code☸️ Kubernetes & Cloud-Native🔭 Obser & SRE🧵 Backend Frameworks & API Tooling🕸️ Networking, Proxies & Service Mesh🏎️ Performance, Load Testing & Chaos🏗️ Platform Engineering & Internal De Platforms🛡️ Security, AppSec & Offensi🦀 Systems Languages, Runtimes & Containers🎨 Ad Frontend & Build Tooling📱 Ad Mobile🚀 2026 Breakouts (Pro Picks)---
🏛️ System Design & Distributed Systems
System Design Primer - The canonical open-source guide to scaling systems, with deep treatment of caching, sharding, and consistency trade-offs.System Design 101 - ByteByteGo's explainers of real-world architectures used as shared in design reviews.System Design (karanpratapsingh) - A book-length treatment of scalability, distributed patterns, and failure modes beyond interview trivia.Awesome Scalability - A curated firehose of engineering blog posts and papers on reliable, performant large-scale systems.Awesome System Design Resources - Concise reference of distributed-systems building blocks and case studies kept current for practitioners.Distributed Systems Topics - A dense index of microservices, scalability, and reliability topics with authoritati.Jepsen - Kyle Kingsbury's framework for distributed database consistency under partitions; the industry standard for correctness testing.PingCAP Talent Plan - Hands-on courses that have you build a real distributed K and SQL layer in Rust/Go.Papers We Love - A community repo of foundational CS papers (consensus, storage, concurrency) for first principles.Awesome Distributed Systems - A tightly curated list of the seminal talks, papers, and books on distributed computing.Awesome Software Architecture - Deep, categorized resources on DDD, CQRS, event sourcing, and clean/hexagonal architecture.Awesome Distributed System Projects - Real open-source distributed systems organized for studying production implementations.🗄️ Databases & Storage Engines
PostgreSQL - The official mirror of the world's most ad open-source RDBMS, the default backend for serious systems.CockroachDB - A distributed SQL database with serializable isolation and sur, prized for global consistency without manual sharding.TiDB - A horizontally scalable HTAP database with MySQL compatibility for mixed transactional/analytical workloads. - The MySQL sharding and clustering system that powers YouTube and Slack.YugabyteDB - A Postgres-compatible distributed SQL database for resilient, geo-distributed deployments.ScyllaDB - A C++ Cassandra-compatible store with a shard-per-core architecture deli predictable low latency.TiKV - A CNCF-graduated distributed transactional key-value store (Raft + Percolator).Apache Cassandra - The battle-tested wide-column store for write-hea workloads.ClickHouse - The fastest open-source columnar OLAP database, the default for real-time analytics o of rows.DuckDB - An in-process analytical database ("SQLite for analytics") ubiquitous in data pipelines.Apache Doris - A high-performance real-time MPP analytical database for sub-second queries.TimescaleDB - A Postgres extension turning it into a scalable time-series database with hypertables.InfluxDB - A purpose-built time-series database and platform for metrics and events at scale.QuestDB - A high-throughput time-series database with SQL and nanosecond timestamps. - A fast, cost-efficient time-series database and long-term Prometheus store.PgBouncer - The lightweight Postgres connection pooler, essentially mandatory for ser and high-concurrency backends.Patroni - The de facto template for Postgres high a and automated failover.RocksDB - Facebook's embeddable LSM-tree key-value engine that serves as the storage layer for countless databases.LevelDB - Google's classic embedded key-value store, still studied as the reference LSM implementation.BadgerDB - A fast pure-Go embeddable LSM key-value store for local persistence.Pebble - CockroachDB's RocksDB-inspired storage engine in Go, a clean reference for modern LSM design.etcd - The distributed, strongly-consistent key-value store that backs Kubernetes.Redis - The ubiquitous in-memory data structure server for caching, queues, and real-time data. - The Linux Foundation's Redis fork that became the community standard after Redis's license change.Dragonfly - A multi-threaded, Redis/Memcached-compatible in-memory store with far higher per-node throughput.libSQL - Turso's open-contribution SQLite fork adding replication and ser.Turso Database - A ground-up SQLite rewrite in Rust with concurrent writes and async I/O.SurrealDB - A multi-model (document/graph/relational) database in Rust designed to collapse the backend data layer.Neo4j - The leading nati database for connected-data workloads like fraud detection and knowledge graphs.🌊 Data Engineering & Pipelines
Apache Airflow - The most widely deployed workflow orchestrator for authoring and scheduling data pipelines.Dagster - An asset-oriented orchestrator with strong typing, testing, and lineage that pros favor o.Prefect - Pythonic, dynamic workflow orchestration with a modern hybrid execution model.dbt Core - The transformation-layer standard for , tested SQL analytics engineering.Apache Spark - The dominant distributed engine for large-scale batch and streaming data processing.Apache Flink - The reference stateful stream-processing engine for low-latency, exactly-once pipelines.Polars - A blazing-fast, multi-threaded Rust DataFrame library increasingly replacing pandas for performance.Apache Arrow - The columnar in-memory format and IPC standard underpinning the modern data stack.dlt - A Python-first, code-driven EL library for loading data into warehouses with schema e.Apache Iceberg - An open table format bringing ACID, schema e, and time travel to data lakes.Delta Lake - A storage layer adding ACID transactions and to lakehouse architectures.Apache Hudi - A lakehouse platform bringing incremental processing and upserts to data lakes.Great Expectations - A framework for , documenting, and profiling data quality.SQLFluff - The dialect-flexible SQL linter and formatter for keeping analytics code clean in CI.Airbyte - An open-source ELT platform with the largest catalog of data connectors.Ibis - A portable Python dataframe API that compiles to 20+ SQL and dataframe backends.Apache Beam - A unified programming model for portable batch and streaming pipelines across runners.PRQL - A pipelined relational query language that compiles to SQL for more maintainable analytics.📡 Message Queues, Streaming & E
Apache Kafka - The dominant distributed e platform and the backbone of most e architectures.Redpanda - A Kafka-API-compatible streaming engine in C++ with no ZooKeeper/J for latency and simplicity.NATS Server - A tiny, ultra-fast messaging system with JetStream persistence, popular for microservice and edge messaging.RabbitMQ - The , protocol-rich message broker that remains a default for reliable task queues.Apache Pulsar - A streaming/messaging platform with separated compute and storage, multi-tenancy, and geo-replication.Temporal - A durable execution platform for writing fault-tolerant, long-running workflows as ordinary code.Cadence - Uber's durable workflow orchestration engine (Temporal's predecessor) for long-running business logic.Apache RocketMQ - A high-throughput, low-latency distributed messaging and streaming platform proven at Alibaba scale.Debezium - The standard change-data-capture platform for streaming database changes into Kafka and beyond.Redpanda Connect (Benthos) - A declarati processor for glue-code-free data plumbing and transformation.RisingWave - A Postgres-compatible streaming database for building real-time materialized views in SQL.NSQ - A realtime distributed messaging platform in Go designed for simplicity, scale, and no SPOF.⚡ LLM Serving & Inference
vLLM - The high-throughput PagedAttention serving engine that is the de facto open-source production inference standard.SGLang - A fast serving framework with RadixAttention prefix caching, prized for structured generation and low latency.llama.cpp - The C/C++ GGUF inference engine enabling efficient CPU/GPU/Apple-silicon local LLM deployment.Ollama - The most popular local model runner, giving a clean CLI/API o.cpp with model management.TensorRT-LLM - NVIDIA's compiler and runtime for squeezing maximum throughput out of Tensor Core GPUs.Text Generation Inference - Hugging Face's production Rust/Python inference server that powered many hosted endpoints.LMDeploy - InternLM's toolkit for compressing and serving LLMs with high-performance TurboMind kernels.LightLLM - A pure-Python, token-attention serving framework favored for lightweight, hackable high-concurrency inference.KTransformers - A framework for running huge MoE models on limited GPU/CPU hardware and kernel injection.Aphrodite Engine - A engine focused on high-throughput serving with broad quantization support.LocalAI - A drop-in OpenAI-compatible API for self-hosted text, image, and audio models across many backends.Text Embeddings Inference - A blazing-fast Rust server for embedding and reranker models, a key RAG-stack component.NVIDIA Dynamo - A datacenter-scale distributed inference framework with disaggregated prefill/decode serving.llamafile - A Mozilla project that packages a model and runtime into a single cross-platform executable.exo - Runs a single large model as a distributed cluster across e devices.🧩 LLM App Frameworks & Orchestration
LangChain - The most widely adopted LLM app framework and integration layer for chains, tools, and retrieval.LlamaIndex - The data framework specializing in ingestion, indexing, and retrieval for production RAG.LangGraph - A graph-based runtime for stateful, human-in-the-loop agent workflows; the enterprise orchestration fa.DSPy - A Stanford framework that programs and auto-optimizes prompts/weights instead of hand-tuning strings.Haystack - deepset's modular, production-grade pipeline framework for RAG and search applications.LiteLLM - A uni proxy/SDK giving one OpenAI-format interface to 100+ model pro with cost tracking.Dify - An open-source, low-code platform for building and operating LLM apps and agents.Langflow - A builder for RAG and multi-agent flows with a Python backend.Flowise - A drag-and-drop UI for assembling LLM orchestration flows.Semantic Kernel - Microsoft's enterprise SDK for planners, plugins, and agent orchestration across .NET/Python.Pydantic AI - A type-safe, Pydantic-nati framework for structured, LLM outputs.Mastra - A TypeScript-first framework for agents, workflows, and RAG, popular in the JS/edge ecosystem.Portkey Gateway - A fast open-source AI gateway with routing, fallbacks, caching, and obser.Guidance - A constrained-generation language for interleaving control flow with token-le.Outlines - A structured generation library guaranteeing JSON-schema/regex/CFG-valid LLM outputs.BAML - A domain-specific language for building reliable, type-safe, testable LLM functions. AI SDK - The TypeScript toolkit for building AI apps and agents across many pro.🤖 AI Agents & Multi-Agent Systems
AutoGen - Microsoft's multi-agent con framework and a research staple.CrewAI - A lean framework for orchestrating role-playing agent "crews" with minimal boilerplate.OpenHands - An autonomous software-engineering agent platform (formerly OpenDevin) that writes, runs, and tests code.MetaGPT - A multi-agent framework that simulates a software company with role-based SOPs.OpenAI Agents SDK - A lightweight, pro framework for multi-agent workflows with handoffs and guardrails.smolagents - Hugging Face's minimalist code-writing agent library.Google ADK - Google's Agent De Kit for building, e, and deploying production agents.CAMEL - A research framework for communicative and role-playing multi-agent systems at scale.Agno - A high-performance framework/runtime for agents with memory, knowledge, and tools (formerly Phidata).SWE-agent - Princeton's benchmark-topping agent that autonomously resol issues.Skyvern - Automates browser workflows using that read pages .Stagehand - An AI browser-automation SDK blending natural language with code for reliable web agents.LiveKit Agents - A framework for building real-time, low-latency voice and multimodal AI agents.ChatDev - A multi-agent software company for collaborative code generation.E2B - Open-source secure cloud sandboxes for running AI-generated code.🔎 RAG, Databases & Agent Memory
Milvus - A cloud-nati database engineered for billion-scale similarity search.Qdrant - A Rust database praised for fast filtered search and simple production operations.Weaviate - A database with strong hybrid search and a built-in module ecosystem.Chroma - A de embedding database that is the default for fast RAG prototyping.pg - The Postgres extension that adds search, the pragmatic default for teams already on Postgres.LanceDB - An embedded, multimodal database built on the columnar Lance format.FAISS - Meta's foundational library for efficient similarity search and clustering.Marqo - An end-to-end search engine with first-class multimodal embedding generation.RAGFlow - An end-to-end RAG engine with deep document understanding and agentic context construction.GraphRAG - Microsoft's knowledge-graph-based RAG approach for global reasoning o corpora.Graphiti - Builds real-time, temporally-aware knowledge graphs for agent memory.cognee - A memory/knowledge-graph layer that replaces ad-hoc RAG with structured pipelines.txtai - An all-in-one embeddings database for semantic search, RAG, and LLM workflows.Unstructured - A preprocessing library that turns PDFs, HTML, and docs into clean, chunked LLM data.Mem0 - A uni memory layer giving agents persistent user/session/agent-scoped memory.Letta - An agent framework (formerly MemGPT) treating memory like an OS with self-managed context.🎛️ Model Fine-Tuning & Training
Transformers - The foundational model library and hub interface underpinning most modern LLM work.Unsloth - Custom Triton kernels that make LoRA/QLoRA fine-tuning 2x+ faster with far less VRAM.Axolotl - A config-dri framework favored for reproducible, production training pipelines.LLaMA-Factory - A unified, GUI-friendly fine-tuning platform supporting 100+ models and every major PEFT method.PEFT - Hugging Face's parameter-efficient fine-tuning library (LoRA, adapters, prompt tuning) used e.TRL - A post-training toolkit unifying SFT, DPO, GRPO, and RLHF trainers on the HF stack.Megatron-LM - NVIDIA's reference for large-scale tensor/pipeline-parallel transformer pretraining.DeepSpeed - Microsoft's ZeRO-based training/inference optimization library for training massi.Accelerate - A thin abstraction for launching PyTorch training across multi-GPU/TPU/mixed-precision setups.Diffusers - The standard toolkit for training and running diffusion models for image/audio/video.OpenRLHF - A Ray-based, high-performance RLHF framework for scalable PPO/GRPO alignment training.verl - Engine's flexible RL post-training library (HybridFlow), a 2026 fa.ms-swift - ModelScope's framework for fine-tuning and deploying 400+ LLMs/MLLMs.XTuner - An efficient, flexible toolkit for fine-tuning LLMs and multimodal models.LitGPT - 20+ from-scratch LLM implementations with pretrain/finetune/deploy recipes.🧪 LLM E, Obser & Guardrails
Langfuse - An open-source LLM engineering platform for tracing, evals, and prompt management; the obser leader.Ragas - A reference-free RAG e library that pioneered faithfulness/relevancy metrics.DeepEval - "Pytest for LLMs" with 50+ metrics and CI/CD-friendly unit testing of model outputs.Phoenix - Arize's OpenTelemetry-nati tracing and e tool for LLM/agent apps.Guardrails - Adds input/output and structured-output guarantees around LLM calls.NeMo Guardrails - NVIDIA's toolkit for adding programmable safety/topical rails to con apps.OpenAI Evals - The widely-used framework and registry for building and running model e.lm-e - EleutherAI's de facto harness for academic LLM benchmarking.promptfoo - A CLI/library for test-dri e and LLM red-teaming/security scanning.lighteval - Hugging Face's lightweight, all-in-one e suite across multiple backends.Giskard - A testing framework that scans ML/LLM systems for , bias, and robustness issues.LLM Guard - A security toolkit for defending against prompt injection and data leakage.OpenLLMetry - OpenTelemetry-based instrumentation for end-to-end LLM app obser.Helicone - An open-source LLM obser proxy for logging, caching, and cost monitoring.Opik - Comet's open-source platform for tracing, e, and monitoring LLM applications.📦 MLOps & Experiment Tracking
MLflow - The de facto open standard for experiment tracking, model registry, and GenAI e.Ray - A distributed compute framework (Train/Tune/Serve/Data) underpinning much large-scale ML infra.BentoML - A framework for packaging models and building/serving production inference APIs.KServe - A Kubernetes-nati model inference platform with autoscaling and vLLM support.ZenML - An extensible MLOps/LLMOps framework for portable, reproducible pipelines across stacks.Weights & Biases - The industry-standard experiment tracking, artifacts, and model management SDK.Aim - An open-source, self-hostable experiment tracker for high- run comparison.DVC - Git-based data and model control for reproducible ML pipelines.Kubeflow - The Kubernetes-native ML platform for pipelines, training, and serving at scale.Flyte - A strongly-typed, Kubernetes-nati orchestrator for scalable ML/data pipelines.Metaflow - Netflix's human-centric framework for building and managing real-world data science projects.Feast - The leading open-source feature store for serving consistent features to training and inference.SkyPilot - Runs training/serving/batch jobs across any cloud or Kubernetes with cost and spot optimization.OpenLLM - Run and deploy open LLMs as OpenAI-compatible endpoints with one command.⚙️ DevOps & Infrastructure-as-Code
Terraform - The dominant multi-cloud IaC tool; fluency here is table stakes for infra engineers.OpenTofu - The Linux Foundation's open-source Terraform fork, increasingly adopted after the BSL relicense.Pulumi - IaC using real languages (TypeScript, Go, Python), where teams want tests, loops, and abstractions.Ansible - The agentless configuration-management and automation engine that remains a fleet-management workhorse.Crossplane - A CNCF-graduated control plane that manages cloud infra as Kubernetes resources.Packer - HashiCorp's tool for building identical machine images across clouds.Vault - The industry-standard secrets management and dynamic-credentials system for zero-trust infra.Consul - Service discovery, service mesh, and distributed KV for connecting and securing services.Nomad - A simple, flexible workload orchestrator scheduling containers and non-container workloads.SOPS - The standard for encrypting secrets in Git (YAML/JSON) with KMS/age.Terragrunt - A thin Terraform wrapper for keeping large multi-en configs DRY.Infracost - Shows cloud cost diffs in pull requests so teams catch expensi changes before merge.LocalStack - A local AWS cloud emulator for de and testing cloud apps without cost or latency.Dagger - A programmable, containerized CI/CD engine that lets you write pipelines as code and run them anywhere.Kubespray - Production-grade Ansible playbooks for deploying self-managed Kubernetes clusters.Salt - An e automation and configuration-management platform for large infrastructures.AWX - The upstream open-source project behind Ansible Automation Platform.Earthly - Reproducible, container-nati that run identically locally and in CI.☸️ Kubernetes & Cloud-Native
Kubernetes - The container orchestration platform that defines cloud-nati.Helm - The de facto Kubernetes package manager for templating and releasing complex applications.Kustomize - Template-free, o Kubernetes configuration management built into kubectl.Argo CD - The leading declarati continuous-delivery controller for Kubernetes.Flux - A CNCF-graduated GitOps toolkit for keeping clusters continuously reconciled with Git.Argo Workflows - A Kubernetes-nati engine for orchestrating parallel jobs and data/ML pipelines.Argo Rollouts - Progressi (canary, blue-green) with automated analysis and rollback.Istio - The most feature-complete service mesh, pro mTLS, traffic management, and telemetry.Linkerd - A lightweight, security-first service mesh for operational simplicity.k3s - A certified lightweight Kubernetes distribution ideal for edge, IoT, and CI.kind - Runs Kubernetes clusters in Docker, the standard for local testing and CI of cluster software.Rancher - A complete platform for managing fleets of Kubernetes clusters across on-prem and cloud.Cluster API - Declarati APIs for managing the lifecycle of clusters themselves.KubeVirt - Runs traditional containers on Kubernetes.Open Policy Agent - A general-purpose policy engine (Rego) for authorization and admission control.Kyverno - A Kubernetes-nati engine using YAML instead of a new language.Harbor - A CNCF-graduated container registry with scanning, signing, and replication.Tekton Pipelines - A Kubernetes-nati framework for building cloud-native CI/CD pipelines.cert-manager - Automates issuance and renewal of TLS certificates in Kubernetes.Kubebuilder - The SDK for building Kubernetes APIs and operators with CRDs.Knati - Ser scale-to-zero and request-dri primitives on Kubernetes.KEDA - E autoscaling on queue depth, Kafka lag, and dozens of other sources. - Backup, restore, and disaster recovery for Kubernetes resources and persistent .Longhorn - A CNCF distributed block storage system for Kubernetes with replication and snapshots.Rook - A storage operator that turns Ceph into self-managing storage on Kubernetes.🔭 Obser & SRE
Prometheus - The de facto cloud-nati and alerting system; its data model defines modern monitoring.Grafana - The uni obser dashboard for metrics, logs, and traces across any backend.Loki - A cost-efficient, label-based log aggregation system that indexes metadata instead of full text.Tempo - A high-scale distributed tracing backend that needs only object storage to run.Mimir - Horizontally scalable, long-term Prometheus storage handling a billion+ acti.Pyroscope - A continuous-profiling database to debug CPU/memory down to the line of code in production.OpenTelemetry Collector - The pipeline for receiving, processing, and exporting all telemetry.Jaeger - The CNCF-graduated distributed tracing system for root-causing latency across microservices.Thanos - Adds global query, unlimited retention, and HA to Prometheus storage. - A blazing-fast Rust obser data pipeline for collecting, transforming, and routing telemetry.Fluent Bit - A lightweight, high-throughput log/metric processor and forwarder for edge and containers.Netdata - Real-time, per-second infrastructure monitoring with zero-config auto-discovery.SigNoz - An OpenTelemetry-nati APM and obser platform (open-source Datadog).OpenObserve - A petabyte-scale obser platform with dramatically lower storage cost.Parca - Always-on eBPF-based continuous profiling to cut CPU and memory usage across a fleet.Pixie - eBPF-powered, no-instrumentation obser for Kubernetes with automatic protocol tracing.Coroot - An eBPF-based obser tool that auto-generates service maps and SLO-dri.Grafana OnCall - Open-source on-call scheduling and incident escalation, an alternative to PagerDuty.Keep - An open-source AIOps and alert-correlation platform to tame alert fatigue.Robusta - Automation and enrichment that adds context and remediation to Kubernetes alerts.🧵 Backend Frameworks & API Tooling
Gin - The most popular Go HTTP framework, for speed and a minimal API.Echo - A high-performance, minimalist Go web framework with strong middleware ergonomics.Fiber - An Express-inspired Go framework built on fasthttp for maximum throughput.chi - A lightweight, idiomatic Go router that composes cleanly with the standard library.gRPC-Go - The Go implementation of gRPC, the default for high-performance internal APIs.Connect-Go - A simpler, browser-compatible RPC framework interoperable with gRPC.sqlc - Generates fully type-safe Go from raw SQL, a fa for teams a hea ORMs.Ent - Facebook's graph-based, schema-as-code ORM for Go with strong static typing.GORM - The full-featured Go ORM that remains the most-used data access layer.golang-migrate - Database schema migrations as a library and CLI for every major database.Tokio - The asynchronous runtime that anchors all production Rust network services.Axum - The Tokio team's ergonomic, modular web framework, now the mainstream Rust choice.Actix Web - An extremely fast, mature Rust web framework topping throughput benchmarks.SQLx - Async, compile-time-checked SQL for Rust with no DSL.Serde - The foundational Rust serialization framework used by essentially every Rust backend.Tower - Composable middleware powering timeouts, retries, and load-shedding in Rust services.Spring Boot - The dominant J for production microservices and enterprise backends.Quarkus - A Kubernetes-native Ja with fast startup and low memory via AOT compilation.Micronaut - A modern J using compile-time DI to avoid reflection.Ktor - JetBrains' asynchronous Kotlin framework for connected server and client applications.Netty - The asynchronous e network framework underpinning much J.NestJS - A structured, TypeScript-first Node framework bringing DI and modular architecture to backends.Fastify - A high-throughput, low-o Node framework with schema-based .Hono - An ultrafast, runtime-agnostic web framework that runs on Workers, Deno, Bun, and Node.tRPC - End-to-end type-safe APIs without codegen, popular in TypeScript monorepos.Prisma - A type-safe ORM and toolkit that has become a default data layer for TypeScript backends.gRPC - The core multi-language RPC framework for efficient, contract-first service communication.Protocol Buffers - Google's language-neutral serialization format, the schema backbone of gRPC.Buf - Modern Protobuf tooling with linting, breaking-change detection, and a schema registry.OpenAPI Specification - The standard for describing REST APIs that dri, docs, and contract testing.🕸️ Networking, Proxies & Service Mesh
Envoy - The CNCF-graduated L7 proxy at the heart of most service meshes and modern edge infra.Traefik - A cloud-nati proxy and ingress with automatic service discovery and Let's Encrypt.Caddy - A Go web ser with automatic HTTPS by default, for operational simplicity.Cilium - eBPF-based networking, security, and obser; the leading modern CNI.Kong - A high-performance, plugin-extensible API gateway built on Nginx/OpenResty.HAProxy - The battle-hardened, ultra-reliable TCP/HTTP load balancer for demanding traffic.CoreDNS - The CNCF-graduated, plugin-based DNS ser service discovery inside Kubernetes.MetalLB - A load-balancer implementation for bare-metal Kubernetes clusters.Pingora - Cloudflare's Rust framework for fast, programmable proxies, serving 40M+ requests/second.Tetragon - eBPF-based runtime security obser and enforcement with low o.Hubble - Cilium's eBPF-powered network, service, and security obser layer.Kubeshark - An API traffic analyzer ("Wireshark for Kubernetes") for real-time network .🏎️ Performance, Load Testing & Chaos
k6 - A de, scriptable load-testing tool (Ja) that has become the modern standard.Locust - Distributed load testing with scenarios defined in Python. - A HTTP load-testing tool for constant-rate attacks with rich reporting.Gatling - A high-performance load-testing framework with an expressive DSL for the J.FlameGraph - Brendan Gregg's flame graph , the canonical way to interpret CPU profiles.bpftrace - A high-le language for Linux eBPF, indispensable for deep production performance analysis.BCC - A toolkit of eBPF-based tools for ad kernel tracing, profiling, and networking obser.async-profiler - A low-o sampling profiler for the J safepoint bias.wrk2 - A constant-throughput wrk fork with corrected latency measurement.hyperfine - A command-line benchmarking tool with statistical rigor for comparing program performance.Chaos Mesh - A CNCF cloud-nati platform for injecting faults into Kubernetes.Litmus - A CNCF chaos-engineering framework for practicing resilience testing.Toxiproxy - Shopify's TCP proxy for deterministically simulating network failures and latency in tests.🏗️ Platform Engineering & Internal De Platforms
Backstage - Spotify's CNCF de portal and software catalog, the de facto standard for IDPs.Kratix - A platform orchestrator that delivers self-service APIs between portals and infrastructure.Score - A platform-agnostic workload specification decoupling de intent from en config.Porter - An internal de platform that deploys apps to your own cloud with a Heroku-like experience.Devtron - An open-source Kubernetes-nati delivery platform unifying CI/CD, GitOps, and security.werf - A CNCF GitOps CLI tying Git, build, and Helm deploy into reproducible delivery pipelines.Kusion - An intent-dri orchestrator for building app-centric internal de platforms.🛡️ Security, AppSec & Offensi
Semgrep - Fast, rule-based static analysis (SAST) with pattern-as-code, widely adopted in CI security gates.TruffleHog - Scans repos, history, and systems for leaked credentials and them live.Gitleaks - A fast, CI-friendly secrets scanner that is a standard pre-commit and pipeline check.Trivy - The all-in-one scanner for container images, IaC, SBOMs, and secrets, ubiquitous in De.Grype - A fast scanner for container images and filesystems, often paired with Syft.Syft - Generates SBOMs (SPDX/CycloneDX) from images and filesystems, foundational to supply-chain security.Cosign - Sigstore's tool for signing and container images and artifacts.Checkov - Policy-as-code static analysis for Terraform, CloudFormation, and Kubernetes.OWASP ZAP - The leading open-source DAST web app scanner and proxy for runtime .Nuclei - Template-driven, high-speed scanning, the standard for automated detection.Subfinder - A fast passi enumeration tool, a staple of recon workflows.httpx - A fast, multi-purpose HTTP toolkit for probing and fingerprinting hosts at scale.naabu - A fast, reliable port scanner in Go built for large-scale recon pipelines.Amass - OWASP's in-depth attack-surface mapping and external asset discovery framework.sqlmap - The definiti SQL injection detection and exploitation tool.Sliver - Bishop Fox's cross-platform ad C2 framework, a modern Cobalt Strike alternative.Havoc - A modern, malleable post-exploitation C2 framework popular in red-team engagements.BloodHound - Graph-based analysis of Acti/Azure attack paths.NetExec - The maintained successor to CrackMapExec for automating network and AD assessments.PayloadsAllTheThings - The go-to reference of payloads and bypass techniques for web app pentesting.SecLists - The essential collection of wordlists for fuzzing, brute-forcing, and discovery.hashcat - The world's fastest GPU-accelerated password recovery tool.Metasploit Framework - The canonical exploitation framework for penetration testing and exploit de.PEASS-ng - The definiti enumeration scripts (linPEAS/winPEAS).Prowler - Multi-cloud security assessments (AWS/Azure/GCP/K8s) mapped to CIS and compliance benchmarks.kube-bench - Checks Kubernetes clusters against the CIS Kubernetes Benchmark.Kubescape - A Kubernetes security platform for misconfiguration, RBAC, and image scanning.Cloud Custodian - A YAML rules engine for cloud go, cost, and security guardrails at scale.ScoutSuite - NCC Group's multi-cloud security auditing tool.Pacu - An open-source AWS exploitation framework for offensi security testing.Wazuh - An open-source SIEM/XDR platform for threat detection, integrity monitoring, and compliance. - Endpoint and digital forensics/incident response at fleet scale.TheHive - A scalable security incident response platform for SOC collaboration.Falco - The CNCF runtime security engine detecting anomalous behavior via eBPF/syscalls.CrowdSec - A collaborati intrusion pre system with crowd-sourced threat intel.OSSF Scorecard - Automatically scores open-source projects' security posture for supply-chain risk.🦀 Systems Languages, Runtimes & Containers
Go - The reference implementation of the Go language and standard library, essential for backend engineers.Rust - The Rust compiler and standard library, the foundation of the memory-safe systems ecosystem.Zig - A simple, explicit systems language and toolchain increasingly used as the build layer under other projects.Awesome Go - The definiti index of Go libraries and tools.Awesome Rust - The canonical curated catalog of Rust libraries, applications, and systems tooling.Go Project Layout - The widely referenced con for structuring real-world Go applications.Uber Go Style Guide - Uber's battle-tested Go style guide adopted by many teams as their standard.fasthttp - A zero-allocation HTTP implementation for Go with order-of-magnitude throughput gains.conc - Sourcegraph's structured-concurrency toolkit for correct, ergonomic goroutine management.Rayon - Data-parallelism for Rust that makes CPU-bound code parallel with a one-line change.clap - The full-featured, ergonomic command-line argument parser powering most Rust CLIs.Firecracker - AWS's Rust microVM tech powering Lambda/Fargate, the reference for secure lightweight .containerd - The CNCF-graduated industry-standard container runtime underpinning Docker and Kubernetes.runc - The OCI reference low-le for spawning containers.Podman - A daemonless, rootless container engine that has become the secure Docker alternative on Linux.CRI-O - A lightweight, Kubernetes-dedicated container runtime implementing the CRI.Wasmtime - The Bytecode Alliance's fast, secure WebAssembly/WASI runtime and Component Model reference.Wasmer - A leading uni WebAssembly runtime for running Wasm across languages and platforms.Bun - An all-in-one JS runtime, bundler, and package manager built for speed.Deno - A secure-by-default, TypeScript-first runtime with a modern standard library.Ruff - Astral's extremely fast Rust-written Python linter and formatter.uv - Astral's ultra-fast Rust package/project manager rapidly replacing pip/Poetry.Biome - A Rust-based, all-in-one formatter and linter positioned as a fast Prettier/ESLint replacement.SWC - A Rust platform for fast TS/JS compilation and transformation used inside many toolchains.typescript-go - Microsoft's native Go port of the TypeScript compiler (the basis of TS 7) targeting ~10x speedups.🎨 Ad Frontend & Build Tooling
React - The core UI library whose Ser and compiler define much of modern frontend.Next.js - The dominant React meta-framework for SSR/SSG/RSC production apps.Svelte - A compiler-based framework (v5 runes) deli minimal-runtime, high-performance UIs.S - The official application framework for building Svelte apps.SolidJS - A fine-grained reacti offering React-like ergonomics with near- performance.Astro - A content-focused framework with islands architecture that ships zero JS by default.Nuxt - The intuitive for hybrid-rendered, full-stack applications.Qwik - A resumable framework eliminating hydration for instant-loading web apps.Vite - The near-uni de and build tool that became the frontend tooling backbone.Turborepo - A Rust-based high-performance build system with caching for JS/TS monorepos.Nx - An extensible monorepo build system with computation caching and rich generators.Rolldown - A Rust-based, Rollup-compatible bundler becoming Vite's unified production bundler.Oxc - A Rust suite of ultra-fast JS/TS tools (parser, linter, resolver, minifier).Rspack - A Rust bundler with a webpack-compatible API for drop-in speed on large codebases.TanStack Query - The standard async-state/data-fetching library for React and other frameworks.TanStack Router - A fully type-safe router with first-class search-param and data-loading APIs.React Router - The routing standard that absorbed Remix into a full-stack React framework.shadcn/ui - The copy-in, unstyled-by-default component collection that redefined design systems.Tailwind CSS - The utility-first CSS framework whose v4 Oxide engine is a modern styling default.📱 Ad Mobile
React Native - The cross-platform native app framework whose New Architecture closed the performance gap.Expo - The recommended production toolchain for React Native with builds, OTA updates, and nati.Compose Multiplatform - JetBrains' declarative UI sharing Compose across Android, iOS, desktop, and web.The Composable Architecture - Point-Free's state-management library for testable, scalable SwiftUI apps.React Nati - The high-performance nati animation library essential to serious RN apps.Now in Android - Google's reference app showcasing modern Android architecture and Compose best practices.Capacitor - A nati for building cross-platform apps with web tech and full native API access.Tamagui - A uni UI kit and optimizing compiler for React Native + web from one codebase.🚀 2026 Breakouts (Pro Picks)
_Repositories that broke out or went viral between late 2025 and mid-2026 - the newest additions to a senior engineer's radar._
OpenClaw - Self-hosted personal AI assistant that wires any LLM into WhatsApp/epo of the year.Codex CLI - OpenAI's Rust-rewritten open-source terminal coding agent (~95k+ stars in 2025).Gemini CLI - Google's open-source terminal AI agent embedding Gemini, one of the most-watched de of 2025-2026.OpenCode - A fast, pro terminal coding agent that became a dominant open CLI agent in 2026.Crush - Charm's polished Go TUI coding agent with multi-model and MCP support.Cline - The leading autonomous VS Code coding agent with human-in-the-loop approval and MCP tools.Kilo Code - An all-in-one agentic VS Code extension merging Roo/Cline ideas, a fast-growing 2025-2026 fork.Goose - Block's extensible on-machine agentic coding assistant, now under Linux Foundation go.Pi (pi-mono) - A radically minimal terminal coding agent + harness toolkit that blew up on a sub-1k-token system prompt.MCP Servers - The reference collection of Model Context Protocol servers, the standard for connecting agents to tools/data.GitHub MCP Server - GitHub's official MCP server for repo/issue/PR/Actions access from agents.Playwright MCP - Microsoft's browser-automation MCP server, the go-to way to gi a real browser.gpt-oss - OpenAI's first open-weight models (120b/20b, Apache-2.0) since GPT-2, a shock release in Aug 2025.DeepSeek-V3.2-Exp - A MoE model introducing DeepSeek Sparse Attention for cheap long-context inference.Kimi-K2 - Moonshot's trillion-parameter MoE built for agentic, long-horizon tasks.Qwen3 - Alibaba's Apache-2.0 dense + MoE family with hybrid thinking modes, a default open-weight backbone.GLM-4.5 - Z.ai's agentic/reasoning/coding foundation models, a top open coding/agent model.Crawl4AI - An LLM-friendly web crawler that hit #1 GitHub trending as RAG data-ingestion demand exploded.Firecrawl - Turns entire websites into LLM-ready Markdown/structured data via a scrape-and-crawl API.Docling - Parses PDFs and office docs into structured formats for GenAI (LF AI & Data).Motia - A unified backend framework merging APIs, jobs, queues, workflows, and AI agents.Better Auth - A comprehensi TypeScript authentication library.Ghostty - A fast, GPU-accelerated, nati terminal emulator.Zed - A high-performance multiplayer code editor from the creators of Atom.---
🤝 Contributing
Found a production-grade repo that belongs here? See CONTRIBUTING.md and open a pull request. Ad, real-world, acti repositories only.
🧭 The Three Lists
| List | Audience | Link |
| ---------------------- | -------------------------------- | ----------------------------------------------------- |
| 150 Essentials | Beginners & Students | awesome-learn-to-code |
| 280+ Comprehensive | E | Main README |
| 300+ Ad | Professionals & Senior Engineers | _(you are here)_ |
<p align="center">
<strong>Built for people who ship. ⚙️</strong><br/>
If a repo here saved you a war-room at 3 a.m., pay it forward with a PR.
</p>
<p align="center">
📄 Licensed under <a href="LICENSE">Apache 2.0</a> · 🧠 Curated by <a href="
<em>If this helped, star ⭐ and share it with your team.</em>
</p>
项目信息
开发语言：未知
Star 数：4
技术要点
该项目属于防护思路方向的开源资源，代码结构清晰，适合安全研究人员学习参考。项目围绕🧠 300+ advanced open-source repos for senior & staff engineers - distributed systems, databases, Kubernetes & cloud-native, observability/SRE, LLM engineering, and offensive/defensive security.展开，提供了可运行的实现，对理解相关安全机制、工具原理与攻防技术具有较高的研究价值。
从技术实现角度看，此类项目的核心在于对底层原理的深入理解与工程化落地。开发者在研究过程中可以结合官方文档与社区讨论，梳理项目的初始化流程、核心模块职责与数据流转，逐步掌握其中的技术细节与设计思路。
使用与学习建议
1. 下载源码压缩包，解压到本地环境。
2. 根据项目类型准备对应的运行环境与依赖。
3. 按照 README 中的步骤完成配置与部署。
4. 在隔离的测试环境中运行验证，避免影响生产系统。
免责声明
本资源来源于开源社区，仅供安全研究、技术学习与合法授权范围内的攻防演练使用，请勿用于任何非法或侵害他人权益的用途。使用本项目产生的一切后果由使用者自行承担。