#technology

Link notes on face-recognition libraries and open-vocabulary object detection.

Short notes on attention, cross-entropy loss, and native vision transformers.

Strategies for keeping K-Nearest Neighbors robust against outliers, from preprocessing and tuning K to weighting neighbors and choosing distance metrics.

Explains how recurrent neural networks process sequences, why they suffer from vanishing gradients, and how LSTM gates fix long-term memory.

Explains how transfer learning reuses a pre-trained source model through fine-tuning for a new task, and why it works so well with limited data.

A summary of the "Attention is All You Need" paper covering the encoder-decoder design, scaled dot-product and multi-head attention, positional encoding, and major Transformer variants.

Clef and Clef-flash are Cloudflare's first self-trained decision models, hosted on Workers AI and open-weight under Apache 2.0. They return a probability for every option of typed noul, choice, and score questions, and are Jev-API compatible.

Six Apache-2.0 decision models from vLLM Semantic Router that answer choice, yes/no, and score questions in one forward pass and return a probability for every option.

Compares full fine-tuning, LoRA, and QLoRA for adapting LLMs, with pros, cons, and guidance on when to use each, plus fine-tuning tools.

Notes on choosing between prompting, RAG, and fine-tuning for LLM applications, covering PEFT, RLHF, chunking, RAG vs CAG, vector databases, and Vietnamese embedding models.

A short introduction to large language models and the three main ways to adapt them: prompt engineering, RAG, and fine-tuning.

Seven practical tips for getting useful results from ChatGPT, from giving context and using variables to splitting tasks, asking for evidence, and building conversation flows.

A taxonomy of prompting techniques with short examples, from zero-shot and few-shot prompting to chain-of-thought, multilingual, multimodal, agent-based, and answer engineering methods.

Eight prompting strategies for LLM applications, including clear instructions, reference text, task decomposition, giving the model time to think, external tools, and systematic evals, plus self-prompting.

Overview of text chunking approaches for RAG (fixed size, recursive, document-based, semantic, agentic), proposition-based retrieval, and the ChunkFormer speech model.

Explains how prompt engineering, RAG, and fine-tuning relate as ways to steer an LLM, with open-source RAG libraries and a comparison of RAG versus cache-augmented generation.

Overview of Meta's Code World Model (CWM), a 32B-parameter language model that simulates code execution, with benchmark results and download links.

How DoorDash benchmarked key-value stores and tuned Redis with hashes, xxHash, protobuf, and Snappy to cut its ML feature store's cost, memory, CPU, and latency.

Explains why combining 3D-CNN with LSTM handles long, multi-stage human actions in video better than either model alone.

Introduces pose estimation, the task of locating human body keypoints in images and video, and the popular YOLO-Pose and MediaPipe Pose solutions.

A debugging story showing how modeling replayed traffic as a Poisson process, rather than fixed intervals, explained a 2x p99 latency gap in GPU model serving.

Step-by-step tutorial on building a RAG system on a Neo4j knowledge graph with LangChain, plus best practices and common pitfalls for GraphRAG.

Overview of recommender system design, from a two-tower four-stage architecture to an AWS Personalize and generative-AI marketing solution, with a comparison of the two approaches.

Explains SENATOR, a framework that uses structural entropy, a knowledge graph, and MCTS to find LLM knowledge gaps and patch them with targeted synthetic-data fine-tuning.

Why forcing strict test-first TDD on a coding agent often costs more than it returns, and what to do instead - outcome feedback loops, regression tests, static-analysis feedback tools, and ATDD on a hexagonal or vertical-slice architecture.

Walkthrough of a TikTok-style two-tower recommender for H&M and an AWS Personalize plus Bedrock ad-copy solution, with a comparison of when to use each.

A first-person account of three AI-assisted workflows (research, retention optimization, and a 30-day content factory) for growing a TikTok channel quickly.

A large curated list of AI tools, ML/DL courses, tutorials, datasets, and miscellaneous bookmarks.

Blog kỹ thuật claude.dev của Anthropic (Anthropic, PBC) cho người xây dựng với Claude, gồm mục trên site, 16 bài trong posts.json, trang /terminal/ và điều khoản site, đã đối chiếu ngày 23/10/2026.

Gym-exercise and media datasets for training and evaluating models.

NVIDIA's synthetic Vietnamese persona dataset built with NeMo Data Designer - 6M personas across 1M records (~5.4 GB), CC BY 4.0, grounded in GSO statistics.

Summary of the Viettel and NVIDIA Vietnamese LLM datasets, covering the curated corpus, the instruction set, and the function-calling benchmark.

Curated courses, tutorials, and articles for learning AI, machine learning, MLOps, and LLMs.

Curated links for learning LLMs, covering prompt engineering, RAG, fine-tuning, LangChain, and model internals.

Pointers to Vast.ai GPU rental and the Awesome Vietnamese NLP resource list.

Curated list of open-source AI frameworks and platforms for infrastructure, fine-tuning, research agents, conversational AI, education, computer vision, edge computing, and finance.

Curated list of sandboxes, runtimes, orchestration platforms, and infrastructure frameworks for running AI agents.

Agent Reach (Panniantong/Agent-Reach, MIT) is a Python CLI that gives coding agents read commands for web and social channels. Checked 3 November 2026, version 1.6.1 matches the latest release, and the PyPI package agent-reach is a different project.

Overview of multi-agent tools and frameworks, covering Make It Heavy, Awesome LLM Apps, CrewAI, and a list of popular AI agent frameworks.

Curated list of agent skills, context-engineering and compression tools, browser automation for agents, and other productivity tools for coding agents.

anymd reads a public URL into structured Markdown, stores it in a private library, and serves it through a URL prefix, REST, CLI, and a remote MCP server.

How the accessibility tree is built, how to keep a UI's tree small and named, and how an agent should prune it before the model sees it.

BrowserOS (browseros-ai/BrowserOS, AGPL-3.0) is a Chromium monorepo that ships BrowserOS neo for agents and BrowserOS for people. Checked 3 November 2026, the newest GitHub release is BrowserOS neo v0.52.0, and neither app was installed.

Claude Ads (AgriciDaniel/claude-ads, MIT) is a Claude Code skill that audits paid-media accounts across 12 platforms from exported evidence, and keeps live account changes behind a mutation gate.

Claude Capsule Kit (CCK) is a hook-based engineering toolkit for Claude Code that adds session memory, dependency analysis, large-file navigation, and parallel crew teams.

Claude Task Master is an editor-agnostic, MCP-based task management system that parses PRDs into tasks and subtasks for AI-driven development.

ClaudeKit is a Claude Code toolkit offering smart guardrails, real-time error prevention, expert subagents, and spec-driven workflow automation.

ClaudeKit CLI (`ck`) is the command-line tool and React web dashboard for managing ClaudeKit projects, with hook diagnostics, a projects registry, and GitHub automation.

ClaudeKit Skills is a collection of 30+ Agent Skills for Claude Code, distributed through a plugin marketplace, covering AI/ML, web development, DevOps, databases, documents, and debugging.

Everything Claude Code (ECC) is a large multi-harness agent performance system bundling agents, skills, hooks, rules, continuous learning, and security scanning.

Side-by-side comparison of seven Claude Code toolkits (ECC, oh-my-claudecode, Task Master, ClaudeKit family, Capsule Kit) by features, architecture, use case, and maturity.

Oh-My-ClaudeCode (OMC) is a teams-first multi-agent orchestration system for Claude Code with staged team pipelines, cross-provider workers, deep interview, and persistent execution modes.

Related Claude Code tooling that does not yet have its own note, starting with the blink-query companion utility.

Cloudflare's open-beta Web Search API sends a query through AI Gateway to Ceramic.ai, Exa, or Linkup and returns one result shape, billed at each provider's list price.

Curated overview of AI coding agents, CLIs, VS Code extensions, IDEs, and app builders such as OpenCode, Devin, Cosine, Dyad, and Kilo.

eve (vercel/eve, npm eve 0.80.1, Apache-2.0, public beta) is Vercel's filesystem-first TypeScript framework for durable backend AI agents - an agent is an agent/ directory of instructions, tools, skills, connections, channels, schedules and subagents; sessions run as Workflow SDK durable workflows with a per-session sandbox, evals, useEveAgent frontend hooks, and Agent Runs tracing on Vercel or self-hosted on Node.

Google's pre-stable agent orchestration runtime. kubectl-style Task, Workspace, and Model manifests, sandboxed as actors on Agent Substrate, with checkpoint suspend and resume.

gstack (garrytan/gstack, MIT) is Garry Tan's skill pack that adds review, QA, and release commands to coding agents. Checked 3 November 2026, main VERSION is 1.93.21.0, the newest GitHub release is v1.92.55.0, and the pack was not installed.

Kanzo (emanuele-em/kanzo, AGPL-3.0, v0.3.0) is a self-hosted kanban board and MCP server in one Rust binary, so a person and coding agents share the same local tasks.

ktx (npm @kaelio/ktx 0.16.0, Kaelio/ktx, Apache-2.0) is a local context layer that turns a SQL warehouse and team docs into a wiki and semantic layer, then serves approved metrics to coding agents over a CLI and MCP; the unscoped npm package ktx and the Khronos KTX texture format are different projects.

Curated list of frameworks and platforms for building multi-agent systems, including Agno, TinyTroupe, Google ADK, Dify, and TradingAgents.

Nuphos (nuphos/nuphos, Apache-2.0, by Zeabur Inc.) is a team workspace for coding agents - Claude Code, Codex, Grok Build, Antigravity and OpenCode run over ACP on local, remote or cloud runtimes, with shared multi-user sessions, team skills, memory and credentials, DevOps GUIs (Kubernetes, Grafana, cloud consoles), Slack/Discord, triggers, and Electron/iOS/CLI clients. Early access; the Desktop app ships PostHog analytics with session replay on by default.

Alibaba's open-source AI code review CLI (ocr) using a hybrid deterministic-pipeline + LLM-agent architecture for precise line-level comments, with delegation mode for host agents and CI/CD integrations.

OpenDuet (CopilotKit/OpenDuet, MIT) is an early self-hosted template for a proactive agent, built on a vendored OpenBot tree, with mail triage and a morning brief, and the server refuses to start without a CopilotKit Intelligence key.

OpenMausBot (milind-soni/OpenMausBot, Apache-2.0) is a local-first chat app that runs a team of coding-agent bots. Checked 3 November 2026, main is package 0.1.139, the newest GitHub release and npm latest are 0.1.138, and the app was not installed.

OpenWorker (andrewyng/openworker, MIT) is a local desktop AI coworker with a Python server and a Tauri shell. Checked 3 November 2026, the Tauri config and Python package on main say 0.3.4, the newest GitHub release is v0.3.1, and PyPI openworker is 0.0.1.

Experimental durable agent harness from Earendil (shipped with Pi 1.0) for long-running, crash-survivable, multi-conversation agent apps over memory/JSONL/SQLite storage, with pluggable extensions, tools, hooks, tasks, compaction, and documents.

pstack (cursor/plugins, MIT, v0.15.36) is Lauren Tan's Cursor plugin of markdown skills. poteto-mode matches one of 23 playbooks and the model follows that file. There is no router process.

REA (npm rea-agents 6.8.0, MIT) is a local CLI and MCP server that returns evidence for native binaries, JavaScript, Electron, .NET, Android, firmware, browsers, and other static formats. Checked 2 November 2026: GitHub lists 420,599 stars, the site is rea.tools, and main has commits after the 25 October 2026 npm release.

Skill Recorder (microsoft/skill-recorder, MIT, Electron + TypeScript) records a screen session on-device, optionally transcribes narration with Whisper and blurs OCR-located text before anything leaves the machine, has GitHub Copilot reconstruct the intent and ordered steps, then renders a SKILL.md into ~/.copilot/skills or a Scout automation. 5.4k stars, v0.7.0.

treg is an open-source tool catalog and credential-injecting proxy for AI agents — one token for 4,000+ priced API endpoints across 121 providers, plus a team's own keys, CLIs and skills, with $1 free credit and 0% markup.

Open-source AI editor copilot plugin for Unreal Engine 5.6 that runs as a dockable in-editor chat tab, drives the editor via Python execution, scene queries, viewport screenshots, and optional MCP/Replicate tools.

Native local-first control plane (Rust, gpui) for driving coding agents such as Claude Code, Codex, Cursor, Devin, Grok, Hermes, Pi, and Antigravity, with optional multi-device sync over Loro and Cloudflare Durable Objects.

Directory of AI SaaS applications for AI workspaces, document handling, research, PDF chat, and document and subtitle translation.

Brave's AI page covers Ask Brave, the Leo browser assistant, and the Brave Search API. The API serves its own index at $5 per 1,000 search requests, with $5 of free credit a month.

Curated tools for analytics and tabular ML, document processing, and knowledge or data management.

Exa is a paid web index and retrieval API for agents. Search starts at $4 per 1,000 requests, contents at $1 per 1,000 pages, and a new account gets $10 in credits each month.

NVIDIA Nemotron Parse 2.0 is an OpenMDW-1.1 vision-encoder-decoder under 1 billion parameters that turns a document image and a task prompt into text, layout classes, and bounding boxes. The Hub card names nvidia/Nemotron-3.5-Parse as the newer model.

SearXNG is an AGPL-3.0 metasearch engine you host yourself. It queries other engines, including Brave, Exa, Google, and Bing, and it does not keep a user profile.

Overview of vector and graph-vector databases, context stores, and RAG frameworks used for semantic retrieval in LLM applications.

Guide to JAX (autodiff, JIT, vectorization, scaling) plus a curated list of LLM application frameworks, agent tooling, and supporting libraries.

Laravel's first-party AI SDK (laravel/ai) — a unified PHP API for agents with tools and structured output, images, audio, transcription, embeddings, reranking, classification, vector stores, and human tool approval.

Tools and models for AI 3D asset generation, text-to-motion animation, and code-driven parametric CAD.

Curated list of AI tools for TTS/ASR, OCR, video and audio download and dubbing, design, and other multimedia content creation and processing.

Curated tools, models, and prompt resources for AI image generation and editing, including text-to-image services, Stable Diffusion, and Midjourney.

Mash Loop (lvt1693/mashloop, MIT) is an endless local jukebox. It uses beat tracking and stem separation to loop each song through seamless jump points and mix a library into one continuous DJ-style set. Weeks old, one author, tested on Windows only.

NovelKit (novelkit.cc) is a Vietnamese long-form fiction studio; the public V2 Lite repo is a local single-operator pipeline under a source-available non-commercial, no-derivatives license, and it is not NovelAI.

SUAPP AI (suapp.ai) is a credit-based cloud design suite for architecture and interiors: AIR rendering, AIC concepts, AIA image-to-video, AIM image-to-3D, AIT assistant, and AIF canvas, run by Metapoleis Limited.

Curated AI video generation tools (text-to-video, lip sync, portrait animation) plus notes on Meta's V-JEPA 2 world model.

Comic-Translate is a desktop app that translates manga, manhwa, BDs and other comics from images, PDF, Epub, CBR and CBZ using detection, OCR, LLM translation and inpainting.

Cotrans is a collaborative online manga translation platform built on manga-image-translator, offering a web UI, browser extension and worker-based backend.

Koharu is a local-first, ML-powered manga translator written in Rust with a Tauri desktop app, combining detection, OCR, inpainting and local or cloud LLMs.

Manga-Image-Translator is the flagship open-source modular manga translation pipeline with pluggable detectors, OCR engines, inpainters and translators, usable via Web UI, CLI, API and WebSocket.

A curated list of additional manga translator projects beyond the main tools, grouped into desktop/web apps, Android apps, libraries, commercial services and discussions.

TachiyomiAT is a Mihon (Tachiyomi) fork for Android that adds automatic manga translation using ML Kit, Google Translate, Gemini or OpenRouter.

Side-by-side comparison of open-source and commercial manga/comic translator tools by stars, license, platform, OCR, inpainting and translation method, with top picks per need.

An evaluation of Claude Code's built-in memory system (CLAUDE.md and auto memory) versus the Claude-Mem plugin, with guidance on when the plugin is worth adding.

Claude-Mem is a persistent memory compression system for Claude Code that captures sessions, compresses them with AI, and injects relevant context into future sessions.

Code Review Graph is a local Tree-sitter knowledge graph for AI coding tools that cuts tokens on code reviews and provides blast-radius impact analysis, with a full install, CLI, and troubleshooting guide.

codebase-memory-mcp (DeusData/CBM) is a zero-dependency, pure C code intelligence MCP server that indexes 162 languages into a queryable knowledge graph.

CodeGraph is an MIT-licensed code knowledge graph with a native Rust kernel, auto-sync, and surgical context retrieval for coding agents.

GitNexus is a zero-server code intelligence engine that indexes codebases into knowledge graphs and exposes them to AI agents through a CLI, MCP tools, and a browser-based Web UI.

A comparison of five memory and knowledge-graph tools (semantica, trustgraph, mempalace, neural-memory, knowns) that combine graph storage with semantic/vector search, with a quick decision table.

Graphify is a Claude Code skill that turns codebases, docs, papers, images, and videos into queryable knowledge graphs, cutting token usage substantially compared with reading raw files.

Knowns is a local-first, self-hostable memory layer for AI-native development that gives AI assistants persistent project context across tasks, docs, memories, and code intelligence.

Mem0 is a popular universal memory layer for AI agents and assistants offering multi-level user, session, and agent memory with Python and Node.js SDKs and an optional managed service.

A feature-by-feature comparison of 15 AI memory, knowledge graph, RAG, and development workflow tools covering storage, memory types, integrations, tech stack, benchmarks, and when each one is worth adopting.

MemPalace is a free, local-only AI memory system that stores raw verbatim exchanges in ChromaDB with a SQLite knowledge graph and reports the highest LongMemEval score at 96.6% R@5.

MemU is a hierarchical memory framework for 24/7 proactive AI agents that cuts LLM token costs and infers user intent without explicit commands, available self-hosted or as a cloud API.

An open-source Python library that gives AI agents persistent, graph-based memory recalled through spreading activation, with MCP tools, a CLI, and offline-first storage.

A curated list of additional memory and knowledge-graph tools for AI coding agents that do not have their own notes.

An open-source Retrieval-Augmented Generation engine that combines deep document understanding, template-based chunking, grounded citations, and agent capabilities.

A codebase intelligence tool that merges a code graph, git history, docs, architectural decisions, and code health into one index for both agents and humans.

An open-source framework for building context graphs and decision-intelligence layers with reasoning engines, temporal queries, provenance tracking, and ontology support.

An evaluation of memory and graph tools for indexing and searching source code, with recommendations by use case and a decision tree.

An agentic skills framework and software development methodology that gives coding agents a complete brainstorm-to-merge workflow built on composable skills.

A CLI that injects structured decision and problem-solving frameworks, with cognitive-bias warnings, into AI prompts for Claude Code, Copilot, and Antigravity.

A self-hosted context development platform that stores, enriches, and retrieves structured knowledge through a multi-model database, graph-native RAG pipelines, and portable context cores.

XERJ (xerj-org/xerj, Apache-2.0, Rust) is a local-first search engine and MCP server for AI agents. One binary auto-indexes folders of code, docs, logs and PDFs into BM25, vector and hybrid search behind an Elasticsearch-compatible API, with an agent memory endpoint. Its headline claim is fewer output tokens for "reference coding". Release candidate (v1.0.0-rc.97), asks users for a field-report PR.

Alibaba's open-source in-process vector database, built on Proxima, offering fast dense and sparse vector search with hybrid filtering and no server to run.

Explains how to use the t-test to decide whether one machine learning model's accuracy is statistically better than another's.

PostClaw is Cassau, LLC's hosted social scheduler at postclaw.fun. The marketing site is Astro 5.18.1, the about page says a Go backend and a React app, and the pricing page lists Free, Connect at $10, Pro at $20, and Business at $50.

Tools for running LLMs and voice models locally, including LM Studio, Ollama, and IchigoLLaMA3.1.

Curated speech tooling covering ASR, TTS, pronunciation assessment, forced alignment, and self-supervised speech models such as HuBERT and wav2vec.

Pointer to Stagehand, an AI browser automation framework for testing and web automation.

ElevenLabs is a hosted AI audio platform (TTS, Instant and Professional Voice Cloning, dubbing, Scribe STT, voice agents, music) with a public API. Vietnamese is in Eleven v4, v3, and Flash v2.5 but not Multilingual v2; commercial use starts at the Starter plan.

V-Clone is a paid, closed-source Vietnamese Windows desktop app sold by Rainbow Music Led (Hải Dương) for voice cloning, film dubbing against SRT/ASS, STT, and offline translation in 12 languages including Vietnamese; the model and licensing details are not published.

ElevenLabs vs VoiceStudio (formerly OmniVoice Studio) vs V-Clone for voice cloning, with Vietnamese support, local vs cloud, sample needs, pricing, licensing, API, and picks for Vietnamese cloning, self-hosting, and commercial use.

VoiceStudio (formerly OmniVoice Studio, debpalash/VoiceStudio, AGPL-3.0) is a local desktop app for zero-shot voice cloning, voice design, dubbing, dictation, and transcription, defaulting to the k2-fsa/OmniVoice model (600+ languages, Vietnamese included). The app code is AGPL; model weight terms are separate.

Notes on an adaptive A/B testing tool architecture that uses Amazon Bedrock and MCP to assign variants based on user context.

Day-by-day summary of Google and Kaggle's free self-paced 5-Day Gen AI Intensive course, covering LLMs, embeddings, agents, domain-specific LLMs, and MLOps.

Learn Harness Engineering is WalkingLab's MIT course on the files, checks, and session rules around a coding agent. The English site has 14 lectures and 8 projects, and the repo was last pushed on 22 October 2026.

Notes on Tang Jie's 2026 Advanced Machine Learning course at Tsinghua, its assignments, and his statements on what comes after the chatbot era.

Summary of Kaggle Grandmaster Chris Deotte's XGBoost experience, from data science foundations and tuning to scaling on GPUs and deploying with refit and cuML FIL.

Notes on Zyphra's ZUNA, a 380M open-source foundation model that denoises, inpaints, and super-resolves scalp EEG by electrode position — a signal layer for non-invasive BCI, not a thought-to-text decoder.

Explains how the Extended Euclidean Algorithm solves the linear Diophantine equation ax + by = c and how RSA uses it to compute modular inverses.

Comparison table of common sorting algorithms with when to use each, stability trade-offs, and Big O complexity.

Guide to choosing an algorithm by its essential properties, finding program errors with well-built test sets, and optimizing a working program.

Covers arrangements, combinations, and permutations, the generation method with lexicographic order, and backtracking with branch and bound for enumerating configurations.

Overview of the main algorithmic problem-solving strategies (brute force, greedy, divide and conquer, dynamic programming, backtracking, and more) and when to apply each.

Short note defining Big O notation as the number of operations relative to input size and how to simplify a complexity function.

Introduces the convex hull problem in computational geometry and lists the common algorithms for computing it.

Survey of linear, non-linear, hash-based, and advanced data structures with their operations, time complexities, use cases, and selection guidance.

Explains how a persistent segment tree keeps every historical version of the data using path copying, and where it is applied.

Explains the segment tree for O(log n) range queries and updates, with example problems (LIS, range minimum) and lazy propagation.

Overview of the Bloom filter probabilistic data structure, its false-positive behavior, and common applications.

Weighted scoring formula for ranking trending posts, with a variable glossary and a worked numeric example.

TypeScript walkthrough of array internals (ArrayBuffer, array lists, ring buffers), linear and binary search, the two crystal balls problem, and bubble, quick, and merge sort.

TypeScript notes on heaps and priority queues, tries, graph representations and traversals, maps, and LRU cache design.

TypeScript implementations of a queue, a stack, and a doubly linked list built on linked nodes, with notes on their trade-offs.

TypeScript examples of recursion covering the base case, the recursive step, and classic recursive problems.

TypeScript notes on tree terminology, traversals, and binary search tree operations.

Reading-progress note for Le Minh Hoang's algorithms book, with links to the notes split out from it.

Curated list of articles, videos, courses, and practice sites for learning algorithms, data structures, and cryptography math.

Overview of caching layers, read/write caching strategies, invalidation approaches, and popular caching tools for reducing latency and system load.

Practical SQL query tuning guide covering execution order, indexing, keyset pagination, prepared statements, CTEs, and a MariaDB slow-pagination case study.

Comparison of MySQL storage engines (MyISAM, InnoDB, MEMORY) and a deep dive into InnoDB pages, rows, row-size limits, and off-page storage formats.

Introduction to graph databases that store data as nodes and edges, with notes on Neo4j, Dgraph, JanusGraph, and TigerGraph.

Survey of NoSQL storage models (document, key-value, wide-column, graph) with representative databases such as Redis, RocksDB, Cassandra, and MongoDB.

Short overview of relational SQL databases including PostgreSQL, MySQL, TiDB, and ClickHouse, with their typical strengths and use cases.

Explains vector databases built to store, manage, and search embeddings via similarity search, and lists their main applications.

GitHub thay endpoint liệt kê stargazers bằng endpoint trả tối đa 30 tuần đếm sao theo ngày, lọc dải created_at trước khi group; bài học chính là response nhỏ hơn không có nghĩa query rẻ hơn.

Tools and approaches for database access in microservices, covering reverse proxies, load balancers, MySQL clustering, and data synchronization.

Summary of how full-text search works (inverted index, stop words, stemming, relevance ranking) and its support in MySQL.

Compares three relational models for storing tree data (adjacency list, closure table, nested set) and their query and update trade-offs.

Explains database sharding across servers versus table partitioning within one database, including goals, drawbacks, and partition types.

Summary and best practices for optimizing a system serving 1M+ concurrent users, covering custom trace IDs, async rate limiting with Kafka and Flink, minimal Go architecture, and community Q&A.

Case study of issuing vouchers at 1M req/s by pre-distributing them into pod RAM, with the trade-offs and the comment Q&A on fraud, Redis HA, and history.

Overview of the algorithms that make Algolia a fast real-time search engine, covering inverted indexes with tries, typo tolerance, ranking, bitset filtering, and index compression.

Explains why a database connection is an expensive OS process, how connection pools work, and how to size a pool using Little's Law, Kingman's formula, and the process-to-core ratio.

Conclusions, best-practice summary, implementation checklist, and community Q&A on a distributed in-memory cache for heavy-read APIs, covering Redis Pub/Sub sync, race conditions, warm-up, and Redis versus Memcached at scale.

Index and summary of a guide to distributed in-memory caching for heavy-read APIs, pointing to the architecture, optimization, and trade-off notes.

Techniques and lessons learned for tuning a local in-memory cache synchronized by Redis Pub/Sub, covering serialization, native HTTP, ETags, eviction, race conditions, and pod lifecycle.

Problem context and reference architecture for serving 1M+ requests per second on heavy-read APIs using local in-memory caches synchronized through Redis Pub/Sub.

Trade-offs, limitations, comparisons with Redis client-side caching, Memcached, and CDNs, and benchmark numbers for a local in-memory cache fed by Redis Pub/Sub.

Case study of how OceanBase handled Alibaba's 11.11 Singles Day peak of 544,000 TPS, covering its architecture, LSM-tree storage, Paxos replication, and trade-offs.

Best practices for running highly available PostgreSQL with Patroni, PgBouncer, and etcd, including when to use the Ivory management UI versus the CLI.

How ProxySQL connection multiplexing lets many frontend connections share few MySQL backend connections, when it is disabled, and how to tune it.

Community-derived best practices for Vitess MySQL sharding, covering cross-shard transactions, CDC and the outbox pattern, distributed transaction patterns, and deployment.

Curated backend engineering resources covering frameworks, libraries, web servers, ID schemes, authorization models, common concurrency bugs, and search, payment, and gateway tools.

Guide to MongoDB, the document-oriented NoSQL database, covering its data model, CRUD operations, indexing, aggregation, replication, sharding, and best practices.

Guide to MySQL, the popular open-source RDBMS, covering storage engines, data types, CRUD, indexing, views, procedures, triggers, replication, and best practices.

PGX (pgrundev/pgx, AGPL-3.0) is a Rust runtime that serves hundreds to thousands of isolated Postgres databases from one process by keeping a single Postgres 18.6-compatible engine, minting databases on connect and branching them with ZFS copy-on-write — built for CI, previews and agent sandboxes, explicitly not production.

Guide to PostgreSQL, the extensible open-source RDBMS, covering data types, JSON/JSONB, arrays, full-text search, index types, window functions, CTEs, and best practices.

Guide to Redis, the in-memory data structure store used as a database, cache, and message broker, covering its use cases, data types, persistence, and best practices.

Overview of Apache Pinot, the distributed real-time OLAP database, its indexing optimizations, and how Uber used it to solve job counting at scale.

Curated list of tools for working with databases, including Docker Compose setups, query builders, LLM-powered query generators, and online SQL playgrounds.

What growth management is — Sean Ellis' growth hacking origins, AARRR vs growth loops, North Star metrics, high-tempo experimentation, growth team structures, and the PMF test.

A curated list of interview experiences, resume advice, practice platforms, courses, and question banks for software engineering interviews.

A short checklist of design principles, programming paradigms, and front-end APIs to review before a technical interview.

What product management is, what a PM actually owns (discovery + delivery), prioritization frameworks, PM vs PO vs PMM vs Growth PM, and the career ladder.

Ten common system design and microservices interview questions with hints on what to clarify and how to structure an answer.

A first-person account of an Amazon SDE-1 interview round, with the mistakes made and lessons on algorithms, system design, and Leadership Principles.

Questions from a fresher backend interview at a unicorn product company, covering Java, data structures, Firebase, networking, and concurrency.

A walkthrough of a senior software developer interview, covering English, coding, system design, project management, version control, refactoring, and adaptability.

Notes for interviewing at a Japanese-owned IT company, including a self-introduction outline, questions to ask HR and the team leader, and cultural observations.

Bogdan Pshonyak's layoff job-search strategy from a Let's Get Rusty video, covering runway and market checks, a specialization niche, resume tailoring with AI, the application funnel arithmetic, and in-person outreach.

A three-round Oracle interview experience in Gandhinagar, India, with the puzzles, Java, SQL, and string problems asked in each round.

Interview outline and technical questions for a Senior Software Engineer (PHP, JavaScript, Laravel 8+) role, plus tips on what interviewers expect.

A curated list of books, online courses, practice websites, and mock interview platforms for studying algorithms and preparing for coding interviews.

Two firsthand accounts of Fresher Java Backend Developer interviews at a fintech company and an outsourcing company, with the OOP, Java core, database, and Spring questions asked.

Common behavioral and background questions for junior or intern web developer interviews, with sample answers and questions to ask the company.

A personal job-search tracker with recruitment portals for outsourcing and product companies in Vietnam, application outcomes, and notes on which companies to apply to or avoid.

Short Q&A reference covering common Laravel interview topics such as CSRF, facades, service container, validation, Eloquent, and caching.

A condensed Q&A list of the top 50 React interview questions covering components, props and state, virtual DOM, Redux, and build tools.

Practical advice on the mindset, timing, and tactics for negotiating salary during tech job interviews, from delaying the number to handling offers.

An interview reference on Vue's KeepAlive component, covering what it does, why it helps, and where to use it.

The AWS 7 Rs migration strategies (rehost, relocate, replatform, refactor, repurchase, retire, retain) and AWS Prescriptive Guidance on large portfolio migrations — move first, modernize later.

Curated reading on HTTPS, SSL/TLS, RFC 7807 problem details and HTTP/1–3 evolution, plus a rundown of the common HTTP methods and their idempotency.

How a home ISP router maps a public dynamic IP and port to a local IP and port through NAT port forwarding.

Explains how AWS VPCs relate to Regions, Availability Zones and subnets, including core VPC components, subnet types and CIDR examples.

A real story of an AWS Lambda bill exploding 70x from a copy-pasted Terraform memory setting, with the cost drivers and lessons for controlling it.

A Java/Spring Boot Dockerfile on Alpine JRE with JVM tuning, graceful shutdown, a curl healthcheck, and OCI labels, balancing image size against maintainability.

A Java Dockerfile that automatically upgrades Gradle plugins and dependencies at build time and uses a compiled Java class as the healthcheck on a distroless runtime.

A Java Dockerfile that builds a minimal custom JRE with jlink and jdeps and runs it on a distroless base image with a busybox wget healthcheck.

An overview of four Dockerfile optimization approaches for Java/Spring Boot applications, compared by image size, security, and performance.

The most heavily optimized Java Dockerfile, combining a hand-picked jlink JRE, Spring Boot layered JARs, tini, SHA256-pinned Alpine, and a non-root CIS-compliant user.

A Python FastAPI Dockerfile on Alpine that strips libraries, bytecode, and unused stdlib modules to get an image under 110MB with a non-root user.

An index of Dockerfile optimization approaches for Python applications, comparing distroless, Alpine, wheel-based offline, and UV plus tini builds.

A simple, maintainable Python Dockerfile using UV on Alpine with tini as init, plus a cross-Dockerfile comparison, best practices, and conclusion for the Python series.

A multi-arch Python Dockerfile that installs dependencies with UV and runs on a distroless Debian 12 base with manually copied shared libraries for high security and small size.

A Python Dockerfile that builds all dependencies into wheels in a builder stage, auto-patches vulnerable versions, and installs them offline in a slim runtime.

A React/Vite Dockerfile that compiles a minimal Nginx from source on Alpine with pre-compressed assets and security headers, plus a comparison and the full Dockerfiles of all five React approaches.

A React/Vite Dockerfile that serves gzip-precompressed static files with BusyBox httpd on the 92.5 KB lipanski/docker-static-website base image.

A React/Vite Dockerfile that compiles a fully static, UPX-compressed Nginx and ships it in a FROM scratch image under 6MB with gzip and Brotli assets.

A React/Vite Dockerfile that embeds the built static files into a single UPX-compressed Go FastHTTP binary and runs it from a scratch image.

An index of five Dockerfile optimization approaches for React SPAs, from simple BusyBox httpd to custom-built Nginx and a Go FastHTTP binary.

A React/Vite Dockerfile that builds a minimal dynamically linked Nginx, copies only its shared libraries into a scratch image, and uses a static C healthcheck binary.

On-Call In Action is Quan Huynh's English SRE book on incidents, alerting, SLOs, and on-call schedules, with Versus Incident as the worked example. The Kindle edition is ASIN B0F679DKJZ, published 17 May 2025.

Bugsink (bugsink/bugsink, Python/Django) là error tracker tự host nhận sự kiện lỗi từ Sentry SDK. Repo tạo 21/08/2024, bản 2.6.2 ngày 22/10/2026, 2145 sao lúc kiểm tra 23/10/2026. Giấy phép phần chính là PolyForm Shield 1.0.0, PolyForm tự mô tả không phải open source, và GitHub API ghi SPDX NOASSERTION.

Curated list of CI/CD tooling covering GitHub Actions, Woodpecker, GitLab CI/CD, Semgrep and SonarQube, plus a sample Node.js-on-EKS pipeline.

Cloudflare's next-generation agentic CLI (cf) covering the entire Cloudflare API via OpenAPI-generated commands, cloudflare.config.ts TypeScript config, and Vite-first local development — the intended successor to Wrangler.

Curated DevOps tooling across GitOps and IaC, systemd and server management, self-hosted deployment platforms, incident management and Cloudflare helpers.

Kubernetes reading list and tools, including Tilt and the Model PVC Provisioning (MPP) approach for shipping large AI models to K8s.

Overview of reverse proxy types and a detailed profile of the Caddy web server, its automatic HTTPS, configuration model, installation and use cases.

Tools for mesh networking and file transfer, a Google File System reading link, and the Cachet status page.

Curated list of self-hosted VPN, proxy and firewall services such as Xray-core, WireGuard, pfSense, SafeLine and headscale, plus Cloudflare Zero Trust.

Guide to fixing the UFW and Docker firewall conflict on a VPS with ufw-docker, plus an overview of the Nixopus VPS management platform.

Open-source workflow and data-pipeline orchestration platforms, covering Kestra and Apache NiFi with their features and requirements.

Notes on a GitOps architecture for managing a fleet of Amazon EKS clusters through a single Kubernetes API using ACK, kro and Argo CD.

Notes on CPU and disk latency, and on OS time slicing and CPU scheduling.

Introduces the Fourier Transform, the Discrete Fourier Transform and FFT, including how epicycles can draw pictures from Fourier coefficients.

A short curated list of tools and SEO reading for producing audio, ideas, background videos and SEO content for Vietnamese websites.

Danh sách 20 engineering blog của công ty product (Netflix, Uber, Meta, Cloudflare, Stripe, Discord, Grab...) chia theo 5 nhóm, kèm URL đã kiểm tra ngày 21/10/2026, bài mới nhất thấy trong feed và những blog chưa xác nhận được.

A curated list of developer tools, algorithm practice platforms, UI icon and color resources, freelance work references, and learning sites for programmers.

Notes from a Grab tech talk on building an API gateway, covering technology choices, design, performance, maintenance and safe upgrades.

A curated list of engineering blogs, newsletters, Viblo author pages, machine learning blogs and other reading sources for software developers, mostly Vietnamese and global tech communities.

JPEG XL for the web: ISO/IEC 18181, VarDCT and Modular, lossless JPEG transcoding, browser support from Chrome 155 and Firefox 158, the picture fallback, and what the October 2026 benchmarks actually show against AVIF.

Experimental platform CSS features — frame-sizing iframes, scroll-marker carousels, text-decoration-inset, overflow auto clip, and scroll-triggered animations via timeline-trigger.

Which file formats to ship for images, video, audio, fonts, and text in 2026, ranked by Chrome's support and Accept-header preference, with fallbacks for Safari and Firefox.

An overview of WebRTC covering its core APIs, signaling, connection setup, NAT traversal, security, libraries and best practices for real-time browser communication.

Viclass plan-time chip on the top-left classroom pill. It counts down the owner's session_duration, and on 1 November 2026 the full-round corner classes are uncommitted while the pane-toolbar commit keeps only the placement comment.

Explains how reflow, repaint and layout shift work in the browser and how to optimize them to keep Cumulative Layout Shift low in Core Web Vitals.

A condensed list of 37 career, coding, communication and mindset tips from a senior frontend developer.

A broad map of the frontend landscape covering Web APIs, architecture models, build tools, performance, design systems, UI libraries, JavaScript libraries and templates.

Notes from David Walsh's interview with a Pornhub web developer on development process, performance monitoring, video player work, tech stack, Web API pain points and team culture.

audiocn (audiocn/ui, MIT, by OrcDev and fortysevenfx) is a shadcn/ui registry of copy-and-own React audio components - about two dozen components (level meters, spectrum and waveform visualizers, faders, knobs, channel strips, a mixer, a player, sound pads), 6 blocks and 18 Web Audio hooks, built on Base UI and Tailwind v4 and themed by shadcn tokens. Not an npm package; installed with npx shadcn add @audiocn/<item>. Launched 2026-10-02.

A curated list of CSS libraries, linters, animation libraries, build tools, zero-runtime CSS-in-JS options, themes and icon sets.

Ely (elygpui.com) is an MIT component library for GPUI, the Rust UI framework behind Zed, not a GPU cloud. It claims about 1,279 components in 44 chapters with a live WebAssembly gallery. It is a weeks-old solo project, git-only, pinned to gpui-pre 0.3.8, and tested only on macOS.

Gr8r Studio (gr8rstudio/gr8r-studio) is a project management UI built with Vite and plain JavaScript ES modules, with no framework and no runtime dependency. It is a front-end-only demo with seeded data in localStorage and a mock login, not a working backend. No license file; three commits, all on 2026-10-07.

Hallmark (Nutlope/hallmark, MIT, v2.0.0, made by Together AI) is a design skill for Claude Code, Codex, Cursor and OpenCode that builds new pages from one of 48 dealt themes under a non-negotiable Floor of design rules, then lints the result in a real browser with sloplint --render and a fresh-context review. 35.7k stars, 1.9k forks, 62 open issues at check time.

Histoire (histoire-dev/histoire, MIT) is a Vite-native story book for developing and documenting components in isolation, the Vite answer to Storybook, led by Guillaume Chau (Akryum). npm latest is 1.0.0-beta.1 (Jan 2026) with Vue 3 and Svelte plugins; main also documents React, Nuxt, SvelteKit, and MCP tools not yet in a release.

Looseleaf (kevintngo/looseleaf, MIT) is a shadcn/ui-compatible React component library where every surface is paper and every line is a seeded SVG drawing — 123 components, 61 blocks and a static shadcn registry, installed as source rather than a dependency.

MyGo (egoist/mygo, MIT) builds desktop apps in pure Go without cgo. Each window shows either a web page in the system webview, with a TypeScript client generated from Go, or native UI drawn by MyGo on the GPU. The latest release is v0.6.2 and it needs Go 1.27+.

NSH (Native Shell, longbridge/nsh, Apache-2.0) is a Rust engine that renders TSX or Rust UI with each platform's own native views, built for AI-generated screens and signed plugins delivered to a host app at run time. Weeks old, v0.1.1.

OrganUI (organui/organui, MIT) is a shadcn-compatible registry of editable React components for healthcare and life sciences. It has 42 three.js organ models, a molecule viewer, a DNA helix, an ECG trace, a dermatome map, a CT/MRI slice viewer and more, with explicit "not for clinical use" disclaimers. Weeks old, v0.4.0.

A curated reference for the React and Next.js ecosystem covering hooks, optimization, design systems, rendering patterns, libraries and a developer roadmap.

SakaniUI (samzydd/Sakaniui, npm @sakaniui/react 0.8.5, MIT) is a single-author React 19 design system built 1:1 from a Figma file - 114+ components, 42 copy-and-edit blocks, three-layer CSS-variable tokens, light/dark plus a Solid/Glass/Liquid surface axis with SVG-refraction liquid glass (Chromium only), Tabler icons and Recharts charts. About 3 months old, 63 stars.

slide.ng (slide-ng/slide.ng, npm @slide-ng/core 1.0.0, MIT, by Erkam Yaman) is a Slidev port for Angular. One slides.md becomes one standalone Angular component per slide, with real Angular template syntax inside, Slidev's layouts, clicks, presenter view, Shiki/Magic Move/Monaco/KaTeX/Mermaid, export to PDF/PNG/PPTX (pictures only)/Markdown, offline builds, themes, an MCP server and an agent skill. Created 2026-10-16, about 100 stars, almost all commits by one author.

TanStack Hotkeys (@tanstack/hotkeys 0.18.0, MIT, TanStack) is a type-safe keyboard-shortcut library for ten frameworks plus vanilla JS - Mod, Vim-style sequences, scopes, app hotkeys, recording, key-state tracking and devtools - with Agent Skills shipped inside the npm packages. Still beta at 0.x.

Text Editor React (HuyPhan1232203/Text-editor-react, MIT, Phan Nhật Huy) is a Word-like TipTap 3 editor built with React 19 - real A4 pages from a published pagination extension, tables split across pages, comments, Yjs-over-WebRTC sharing with the key in the URL fragment, and PDF export through the print dialog or a Puppeteer backend.

Tips, libraries, UI kits, testing tools and templates for the Vue and Nuxt ecosystem.

Vue TermUI (vue-terminal/vue-termui, MIT, posva) is a Vue 3 custom renderer for terminal apps on OpenTUI's native Zig engine — flexbox layout, forms, focus, and a Vite plugin with in-terminal HMR; ground-up rewrite from the old Yoga-based stack. npm vue-termui 0.3.0.

A curated list of newer frameworks, toolchains and scaffolding tools for web (frontend and backend) and desktop apps.

Design, performance, and monitoring best practices for building reliable and secure IoT systems.

Overview of short-range, long-range, and wired connectivity protocols used in IoT and when to choose each.

Comparison of payload serialization formats for IoT, from human-readable JSON to compact binary formats like CBOR.

Four-stage IoT development workflow covering prototyping, integration, deployment, and maintenance.

Benefits, device types, and use cases of processing IoT data at the edge instead of in the cloud.

Overview of microcontrollers and single-board computers commonly used for IoT projects and what each is best suited for.

Comparison of application-layer messaging protocols for IoT (MQTT, CoAP, HTTP, AMQP) with their features and best-fit scenarios.

Overview of smart home systems, covering components, protocols, use cases, automation scenarios, security, and popular platforms.

Key industrial IoT (IIoT) use cases such as predictive maintenance, asset tracking, quality control, and supply chain optimization.

Notes on anyon_e, Byran Huang's MIT-licensed open-source 13.3 inch laptop built around a FriendlyElec CM3588 (RK3588) module, a Samsung 4K AMOLED panel, and a custom power board, keyboard, and CNC chassis.

Jumper (KingKongRobotics/jumper, Apache-2.0) is a 22-joint crab robot from KingKong Robotics, with an open toolkit for skins, MuJoCo motion training, and scenes.

Curated list of visual programming tools and cloud simulators for prototyping and simulating IoT devices and flows.

Curated list of MQTT clients and debugging tools for testing IoT messaging, firmware, and network traffic.

Curated list of firmware development tools and edge AI projects for building IoT devices.

Gokino (gokino.dev) is a browser Arduino IDE that compiles sketches in the cloud, flashes boards over WebSerial, and meters an AI assistant with credits.

Learn Arduino (learn-arduino-red.vercel.app) is Dilum Sanjaya's free browser-based 3D course — 16 pre-wired Arduino Uno lessons with adjustable parameters, simulated code, serial output and a schematic view, in 13 languages, with no backend or account.

Notes comparing how asynchronous execution works in JavaScript (call stack and callback queue) and in PHP (Fibers as green threads).

Notes on the Go runtime scheduler and the operating-system scheduler article it builds on.

Explains how browsers execute inline, `defer`, `async`, and module scripts, and when to choose each for page performance and correct execution order.

How a .class file becomes a running application: loading and class identity, linking and verification, initialization, interpreter/JIT/AOT, threads and happens-before, heap allocation and GC roots, and JVMTI/JMX/JFR monitoring.

Explains layered package design in Go, why it follows from the ban on circular imports, and a prioritized list of refactorings for breaking circular dependencies.

A Go race condition is shared state whose result depends on unsynchronized interleaving; a data race is the memory-access case, and the usual fixes are a small mutex, an atomic, or a single owner behind a channel, checked with go test -race.

Curriculum outline of Viet Tran's scalable Golang course, covering language features, APIs, databases, async jobs, deployment, gRPC, microservices, and DevOps.

Why ExecutorService fan-out leaks threads and waits too long, and how StructuredTaskScope models subtasks as one unit: fork, join and close, joiner policies, timeout semantics, interrupt-based cancellation, JSON thread dumps, and the JDK 27 preview / JDK 28 final status.

Pointer to MicroProfile, an API library that provides tooling for building microservices in Java.

Curated list of C and C++ libraries, currently featuring the Clay high-performance UI layout library.

Notes on dependency injection for Android with Dagger (components, modules, scopes) and a Koin pitfall with generic types.

Curated Spring Boot resources including a tutorial series, Dockerizing a Spring Boot app, Spring Authorization Server, and the MapStruct DTO mapper.

Hub of JavaScript and TypeScript notes covering language fundamentals, the event loop, performance, the type system, tooling, libraries, and interview prep.

Laravel 13.36 (framework PR 61833, released 2026-10-13) adds a Rule:: builder for every validation rule that was still string-only (Rule::requiredWithout, Rule::url()->protocols, Rule::uuid()->version, Rule::distinct()->ignoreCase, and others), plus Rule::date()->equals. Builders give IDE autocompletion and pass parameters as arrays, so values containing commas, quotes or pipes no longer break; string rules still work and can be mixed in.

Curated list of Node.js web frameworks, ORMs, libraries, an experimental low-latency runtime, and a request validator.

Rul (rul-lang/rul, .rl) is a dependency-free Rust engine in which an AI agent emits a 128-bit-per-node SSA dataflow graph of f64 columnar ops, a single forward pass verifies it, three execution tiers run it, and a 991-line emitter writes a standalone Linux x86-64 ELF with no LLVM or libc. The 192 B program and 817 B ELF are real; the 5.78 ns/node and 6.38 µs headline figures did not reproduce (19.4 ns/node, 19.7 µs on Apple silicon), and the 'language' has 15 opcodes, no control flow, and a stub I/O op.

tsc-ts (vercel-labs/tsc-ts, npm tsc-ts 0.0.5, Apache-2.0) is a Vercel Labs experiment that ports the TypeScript 7.0.2 compiler (typescript-go / tsgo) back into TypeScript and compiles it to a native executable with scriptc. It takes the same arguments as tsc, adds a stdin/stdout --api server and a tsc-ts/api programmatic API, and claims its scanner, parser, binder, checker, emit, and build modes match tsgo on TypeScript's test corpus.

A checklist of security measures for login, logout, forgot-password and registration flows, plus practical deployment caveats.

ptn1411/skill ("Authorized Artifact Auditor", Python, no license file) is a pack of about 35 agent skills plus Python CLIs for reverse engineering and auditing software you own - fingerprinting, unpacking, decompiling, web and network recon, SBOM, CVE lookup. The README and MASTER_POLICY.md say defensive use only, but several skills in the repo (master-unlock, dotnet-keygen, binary-patcher, frida-hooker) are written as bypass and keygen tools. Read before installing.

alexgaoth/awesome-reverse-engineered is a searchable index of about 12,900 open-source projects that exist because someone reverse engineered a closed app, API, device, protocol, file format, game or AI tool, with a website, a JSON dataset (CC0) and an agent skill that searches it. Use it to check whether the work is already done before reverse engineering something yourself.

RedAmon (samugit83/redamon, MIT) is a self-hosted Docker penetration-testing framework that maps an authorized target into a Neo4j graph, runs an agent from a Kali sandbox behind approval gates, and can open a GitHub pull request for a fix.

security-audit (cloudflare/security-audit-skill, MIT) is an agent skill that runs a six-phase defensive security audit - reconnaissance with a coverage ledger, isolated hunters, adversarial candidate validation, schema-validated findings.json, independent record verification, and target-neutral reports - seeded from Cloudflare's fleet-wide vulnerability discovery harness. 48.2k stars, 3k forks, 98 open issues at check time.

A curated list of tools for secret management, vulnerability scanning, cryptography, security analytics, pentesting and PII detection.

Notes on Alfonso Campodonico's "Building your own sandboxes" (2026-10-18) - vmkit runs Firecracker or Cloud Hypervisor microVMs behind one API, each VMM inside a rootless namespace jail with a default-deny per-VM network, and kiln turns OCI images into one-erofs-per-layer microVM images. Main lesson - the VMM is the easy part, the policy around it is the work.

Analysis of the axios vulnerability chain (prototype pollution → CRLF injection → request smuggling) that bypasses AWS IMDSv2 and steals IAM credentials.

A checklist of clean code rules covering naming, functions, comments, source structure, objects, tests, and common code smells.

Clarifies concurrency, parallelism, asynchronous execution, and multithreading, and explains why event-loop async beats threads for I/O-bound work.

How to design a multi-service system around failure: timeouts that mean unknown, partial failure, clock skew, replication lag, sagas and compensation, duplicate or out-of-order messages, retry budgets, tracing, and the trade-offs behind each choice.

Explains the core problems of distributed systems: partial failures, unreliable networks and clocks, synchronous versus asynchronous networks, and long-tail latency.

When a modular monolith is enough and when splitting into services pays off: boundary enforcement, in-process calls versus network calls, the Orders–Billing example, Spring Modulith and ArchUnit, and a Strangler Fig migration path.

Applies the Universal Scalability Law and Little's Law to scaling teams, with lessons on delegation, communication overhead, and limiting work in progress.

Explains the 3-2-1 data backup rule (three copies, two media types, one off-site), how to implement it, the questions to ask when choosing a solution, and related strategies such as 3-2-1-1-0 and 4-3-2.

Explains the hostname routing pattern, which isolates API services by giving each one its own hostname, with typical use cases and trade-offs.

Explains the HTTP header routing pattern, where a request header selects the target service, action, version, or feature, along with its pros and cons.

Explains the path routing pattern, which groups APIs under one hostname and separates services by URI, with NGINX, API Gateway, and CloudFront implementations on AWS.

Describes the Circuit Breaker pattern, which stops a caller from repeatedly calling a failing or slow service, its three states, and how to build it on AWS.

Describes the Event Sourcing pattern, which stores every state-changing event in an immutable event store and rebuilds state by replay, with challenges and AWS implementations.

Describes the Hexagonal Architecture (Ports and Adapters) pattern for loosely coupled, independently testable applications, including how to apply it on AWS Lambda.

Describes the Retry with Backoff pattern, which automatically retries operations that failed from transient errors using exponential backoff and jitter, with AWS implementations.

Describes the Saga Choreography pattern, which keeps data consistent across services through event subscriptions instead of a central coordinator, with an AWS order-flow example.

Describes the Saga Orchestration pattern, which uses a central orchestrator such as AWS Step Functions to coordinate distributed transactions and compensating actions.

Describes the Strangler Fig pattern for incrementally migrating a monolith to microservices behind a proxy layer, with AWS implementations, benefits, and challenges.

Describes the Transactional Outbox pattern, which solves the dual-write problem by saving events in the same transaction as the data, using an outbox table or change data capture.

Explains how an ambassador helper service acts as an out-of-process proxy that sends network requests on behalf of a client, adding resiliency, routing, and security without changing the application.

Describes how a façade or adapter layer translates between subsystems with different semantics so a new application's design is not constrained by legacy or external systems.

Explains how to decouple long-running backend processing from the frontend using HTTP 202 and a polling status endpoint, with an Azure Functions implementation outline.

Describes creating a separate backend service per client interface so each frontend gets a tailored experience instead of forcing one shared backend to serve conflicting needs.

Explains how isolating application components into separate resource pools keeps a failure or overload in one pool from cascading, with a Kubernetes resource-limit example.

Explains how to load data into a cache on demand from a data store, and how to invalidate entries on update to keep the cache consistent, with a .NET example.

Describes decentralizing workflow logic across services that react to each other's events through a message broker, instead of relying on a central orchestrator.

Explains how a circuit-breaker proxy with Closed, Open, and Half-Open states stops an application from repeatedly calling a failing remote service and prevents cascading failures.

Explains how to move large payloads out of a messaging system by storing them externally and sending only a claim-check token, with Azure Blob Storage and Service Bus examples.

A catalog of Azure cloud design patterns with one-line summaries and the Well-Architected pillars each one supports, plus the distributed-computing fallacies they address.

Explains how to undo the steps of an eventually consistent, multi-step operation by running compensating actions instead of a traditional ACID rollback, with a travel-booking example.

Explains how multiple consumers can process messages from a single queue concurrently to improve throughput, scalability, and availability, including poison-message and idempotency concerns.

Describes consolidating multiple tasks into a single computational unit to raise resource utilization and cut cost, and the scalability, lifetime, security, and contention trade-offs involved.

Covers Command Query Responsibility Segregation, which splits read and write operations into separate models to improve performance, scalability, and security.

Describes deploying multiple independent copies of a scale unit (stamps) to host different tenants, gaining near-linear scale, limited blast radius, and regional data placement.

Describes how to manage and distribute configuration to edge workloads that must keep running offline, comparing an external configuration controller with an internal configuration provider.

Explains storing every change to an entity as an immutable sequence of events in an append-only store, with materialized views, replay, snapshots, and trade-offs.

Explains moving configuration out of the deployment package into a centralized store to avoid redeploys and share settings, with Azure App Configuration and Key Vault options.

Explains delegating authentication to an external identity provider using claims-based tokens for single sign-on, less admin overhead, and stronger security, with Microsoft Entra ID examples.

Explains protecting services with a dedicated, low-privilege broker that validates and sanitizes requests before handing them to a trusted host, shrinking the attack surface.

Explains using a gateway to combine multiple client requests into one, cutting chattiness and latency between clients and backend services.

Explains offloading shared cross-cutting functions such as SSL termination, authentication, and throttling to a gateway proxy so individual services stay simpler.

Explains routing requests to multiple services or service versions through a single layer-7 gateway endpoint, enabling backend abstraction and zero-downtime deployments.

Explains deploying self-contained backend nodes across geographic regions in an active-active setup so any node can serve any request, cutting latency and surviving regional outages.

Explains exposing application health-check endpoints that external monitoring tools poll to verify dependencies, response codes, and latency, plus security and caching considerations.

Explains emulating secondary indexes in NoSQL stores with separate index tables, comparing denormalized, normalized, and partially normalized layouts and their consistency costs.

Explains coordinating distributed instances by electing one leader via a distributed mutex or consensus algorithm, including heartbeat, lease, and failure considerations.

Explains generating prepopulated, read-only views over source data to speed up complex queries, with event-driven or scheduled refresh and eventual-consistency trade-offs.

Explains building a bridge component that relays messages between incompatible messaging infrastructures, such as MSMQ and Azure Service Bus, without changing existing systems.

Reference implementation that combines global routing, gateway offloading, and health endpoint monitoring to give HTTP/HTTPS applications a secure entry point on Azure.

Decompose a complex processing task into a chain of reusable, independently deployable and scalable filters connected by pipes.

Let a system process high-priority requests ahead of low-priority ones, using a priority-aware queue or separate queues per priority level.

Announce events to many interested consumers asynchronously through a message broker, without tightly coupling senders to receivers.

Validate third-party software artifacts in an isolated pipeline before they are trusted and released into the internal supply chain.

Place a queue between a task and the service it calls so the queue buffers intermittent heavy loads and smooths out peak demand.

Control resource consumption by capping the rate of access to a capacity-limited service, avoiding throttling errors and keeping throughput predictable.

Handle transient failures when connecting to services or network resources by transparently retrying failed operations, with backoff and idempotency in mind.

Keep data consistent across microservices with a sequence of local transactions and compensating transactions, using choreography or orchestration instead of two-phase commit.

Coordinate a set of distributed actions as a single operation, with a scheduler, agents, and a supervisor that recovers or compensates failed steps.

Process related messages in a defined order without blocking other message groups, using categories and message sessions to keep FIFO while scaling out.

Split a data store into horizontal partitions (shards) using lookup, range, or hash strategies to improve scalability and performance.

Deploy supporting components such as logging, monitoring, and proxies in a separate process or container attached to the main application, sharing its lifecycle.

Serve static assets from a cloud storage service, typically fronted by a CDN, instead of from application servers to cut cost and improve performance.

Incrementally migrate a legacy system by routing functionality through a façade and replacing it piece by piece with new services until the old system can be retired.

Control resource consumption per application instance, tenant, or service so the system keeps meeting its SLAs under sudden demand spikes.

Give clients a short-lived, scoped token for direct access to a storage resource so data transfer bypasses the application server.

Explains how to automatically confirm bank-transfer payments for orders by polling transaction history and matching a transfer-content syntax, and compares third-party services, direct bank APIs, and non-public APIs.

Two real-world case studies showing how tiny changes, such as reordering one line of code and moving CORS headers to Cloudflare, cut CPU by 70% and data transfer by 63% on high-traffic systems.

Covers foundational debugging principles, advanced techniques for memory and concurrency problems, common debuggers, profilers and analyzers, and best practices for systematic debugging.

Practical advice on choosing a compression format and settings (RAR, 7z/LZMA2, zip, gzip/brotli, zstd, dictionary size, solid mode) based on file type and speed, ratio, and compatibility needs.

Explains how hedged requests reduce long-tail latency in distributed systems by sending a duplicate request to another replica after a p95 delay, with the trade-offs around idempotency and choosing the delay.

A system design walkthrough of an Instagram-like photo-sharing service covering requirements, capacity estimates, database design, newsfeed generation with hybrid push/pull, caching, and key generation.

Notes from a Golang Vietnam thread on Kafka consumer DLQ and retry design, covering which errors to retry, where delay lives, when to replay or drop, ordering, backpressure, and delivery semantics.

Explains how the Outbox Pattern with Change Data Capture (Debezium and Kafka) solves the dual-write problem in microservices, including the outbox table structure, duplicate handling, and eventual consistency.

Covers three ways to protect a system from overload and p99 degradation - gateway request body limits (with benchmark results), load shedding with overload detection methods, and hotspot mitigation via small caches and request collapsing.

A six-step evolution from a single file-storage server to a scalable backend: shared metadata, load balancing and autoscaling, service boundaries and gateway auth, direct upload with presigned URLs, asynchronous events, and caching, CDN, and rate limiting.

An overview of how a search engine works, covering its crawler, indexer, query processor and ranker components, the request flow, common ranking algorithms, and key challenges.

An index of PDF references and diagrams attached to the system design notes, from TinyURL designs to VietQR and Discord compression charts.

Community discussion on storing document edit history using versioned diffs with periodic full snapshots, plus comparisons to persistent segment trees, ropes, and git blobs.

An overview of system design fundamentals: the CAP consistency-availability trade-off, eight common problems with standard solutions, and a curated list of learning references.

Defines non-functional requirements and capabilities, and lists the steps to gather requirements before building an application.

A table of contents for the system design notes, covering fundamentals, software architecture, technical solutions, resilience patterns, and attachments.

A pointer to a sample backend blog application on GitHub for studying application architecture.

Covers retry with gradual backoff, the Circuit Breaker pattern, distributed transaction techniques such as SAGA, and the Outbox pattern for reliable messaging.

A survey of software architecture patterns (Clean, Hexagonal, Microservices, EDA, DDD, CQRS, LMAX, MVC family) and design principles (SOLID, DRY, AOP, Command Bus), with references and examples.

A curated catalog of detailed technical solutions and case studies, from JWT revocation, caching, and background jobs to TinyURL, ID generation, rate limiting, migrations, and scaling for traffic spikes.

Glossary of concepts and formulas for a workspace target and points system, such as goals, scenes, outputs, achievements, and working-day adjusted goals.

Walks through a system design interview for a task scheduler, covering requirements, scalability with message queues and priority queues, fault tolerance with retries and backoff, core components, and design trade-offs.

Compares solutions to the Top K heavy hitters problem at 10 billion views per day (Redis sorted sets, Count-Min Sketch, OLAP pipelines, batch jobs, SQL) and recommends a hybrid fast path plus slow path architecture.

Summarizes how YouTube scaled its backend to serve billions of users by putting Vitess on top of MySQL to enable horizontal scaling.

A guide for PMs and delivery managers on balancing profit and loss against software quality, with case studies, team-mix ratios, metrics, and a planning checklist.

A short reading list on product management and customer experience from Ngoc Hieu's blog.

Curated articles on Agile/Scrum practices such as backlog grooming and estimation, plus an introduction to OKRs.

200Lab's free browser game at systemdesign.200lab.io. You place components, the server simulates the traffic, and a level passes only inside its success, latency, and budget limits, with up to three stars for a cheaper design.

sys0.in is Rakesh K's system design course, still being written: a public syllabus of 19 modules, two lessons and two labs marked free, and a yearly founding membership. Checked 2 November 2026 from this network: the founding year was shown as ₫1,408,525, opening a lesson still asked for a Google sign-in, and payments go through Dodo Payments.

Syscraft is Nguyen Thanh Minh's hosted system design lab at syscraft.app: draw an architecture, run simulated traffic, and pass graded challenges. Checked 2 November 2026 without an account: the marketing page lists 7 topics and Pro at $9 a month through Lemon Squeezy, while the learn index lists 42 lessons and the challenge index lists 19.

MIT discrete-event system design simulator at breakscale.tech (workersio/breakscale). The browser, the VS Code webview, and the MCP server share one engine: 33 node kinds, 23 presets, and four challenges. Checked 2 November 2026: 1,953 stars, share links seal designs with AES-GCM and keep the key in the URL fragment, and the website also loads Vercel Analytics.

Links to software architecture documentation tools and frameworks: the C4 model, Structurizr, and arc42.

The Future of Software Testing is Katalon's manifesto and discussion series at thefutureofsoftwaretesting.com. Checked 2 November 2026: four named contributors, a September 2026 manifesto, two published articles, two titles still upcoming, and a how-to track marked coming soon.

e2e (npm e2e 0.18.0, tester-army/e2e, Apache-2.0) is TesterArmy's pre-1.0 web and mobile runner that mixes natural-language agent steps with locator assertions, using Playwright in the browser and agent-device on devices; npm versions 0.0.x are the older willscott/e2e OpenPGP library.

A curated list of tools for API and integration testing, load testing, architecture checks, browser testing and dependency analysis.

Introduction to and comparison of the Trunk-based Development and Forking Workflow models for managing a Git project.

Two handy websites: Radio Garden for listening to live radio worldwide and VERT.sh for private online file conversion.

Useful Windows settings and hidden features, from faster sign-in and dark mode to clipboard history, Night Light and handy shortcuts.

RoboHash, DiceBear, and Peek turn a string into a stable avatar. RoboHash composites licensed artwork over HTTP or Python, published DiceBear is 10.7 with 61 styles while 11 is still a release candidate, and Peek is a small MIT SVG face with an optional React 19 component.

Local dev environment managers, benchmarks, debugging, monitoring and other software development tools.

Crow's foot notation reference and a list of tools for drawing architecture and database diagrams.

Git and GitHub tools, self-hosted servers, hooks and a guide to rewriting history with git filter-repo.

Insomnia (krishhgg/Insomnia, MIT, Swift) is a macOS 26+ menu bar app, not the Kong Insomnia API client, that keeps a MacBook awake with the lid closed for a timed session (e.g. long AI coding-agent runs) via pmset disablesleep, with lid-close actions, battery/heat floors and a launchd recovery backstop. Source review found no telemetry; its main risk is a passwordless sudoers rule for four pmset commands usable by any process running as the user.

Resources for sound effects, GIFs, photo editing and PDF processing with Stirling PDF.

Ondine (wevm/ondine, Rust) biên dịch sơ đồ Mermaid thành SVG từ trước, kèm CLI, plugin Vite cho file .mmd, ondine/remark và WASM tùy chọn. Repo tạo 18/10/2026, npm 0.0.3, 371 sao lúc kiểm tra 23/10/2026. LICENSE trên main là MIT của weth, LLC và ghi phần port từ Pele; LICENSE trong gói npm 0.0.3 không ghi tên chủ bản quyền.

A curated collection of free web tools for image editing, diagrams, downloads, conversion, privacy, developer utilities and wallpapers.

openGym (DuarteSantos8/openGym, AGPL-3.0) is a self-hosted workout and body-weight tracker with passkeys, offline sync, and an optional bring-your-own AI coach; its exercise images are third-party and are not under that license.

Pele (obsidianmd/pele, MIT, by kepano) renders Mermaid diagrams to SVG with a small synchronous TypeScript library that has no runtime dependencies and themes through CSS variables. It targets Mermaid 12.1.0 syntax with its own layout and look. Weeks old, 1.8.0.

ToolCheckMac (huytran1120/ToolCheckMac, MIT) is a native SwiftUI app for checking a used Mac before buying - battery, SSD SMART, MDM/DEP, Activation Lock, serial cross-check, interactive keyboard/screen/audio/camera/trackpad tests and a 0-100 score with PNG/PDF report. Source review found no network code, no sudo and read-only system commands; release builds are unsigned and not notarized.

Open-source business applications and a short list of terminal user-interface clients.