2026-05-21

Denoise

The AI agent ecosystem is rapidly standardizing around "skills" as the new unit of functionality, with a strong focus on Anthropic's Claude.

Today's trend is not about individual agents, but the emergence of a "skill store" model where developers package and share specific agent capabilities.

2026-05-212026-05-21T15:46:27Zrules v1Healthyrepos 41signals 41

Top 3 changes

  • anthropics/claude-plugins-official shows a platform vendor formalizing its agent ecosystem by creating a central directory for skills.
  • colbymchenry/codegraph shows the push for local, optimized tooling to reduce LLM token consumption for coding agents.
  • CloakHQ/CloakBrowser shows persistent developer interest in anti-bot detection and web automation tools, a parallel and durable trend.

Strategic insights

#01The "Skill Store" model is consolidating. The volume of repos named "skills" (e.g., `dotnet/skills`, `mattpocock/skills`) and Anthropic's official `claude-plugins-official` repo indicates a market convergence on reusable agent components.
#02Anthropic's Claude is exhibiting strong ecosystem gravity. Multiple new and rising projects like `ruvnet/ruflo` and `multica-ai/andrej-karpathy-skills` are explicitly built for or tailored to Claude, suggesting significant developer mindshare.
#03"AgentOps" is the next layer of abstraction. Tools like `multica-ai/multica` and `ruvnet/ruflo` are moving beyond building single agents to focus on orchestration and management, signaling maturation in the agent development lifecycle.
#04A tension exists between cloud-native agents and local optimization. While the ecosystem coalesces around cloud APIs, projects like `colbymchenry/codegraph` and `decolua/9router` highlight the developer drive to control costs and data locally.
#05Perennial 'grey hat' tooling remains a constant trend. Projects for anti-bot detection (`CloakHQ/CloakBrowser`), OSINT (`soxoj/maigret`), and censorship circumvention (`alireza0/s-ui`) consistently appear, representing a stable user base.

Trend tracking

Today

Today's trend is dominated by newly surfaced projects, led by the knowledge compilation `trimstray/the-book-of-secret-knowledge` and the AI tool `multica-ai/multica`.

7 Strengthening · 4 Continuous · 8 First seen
Expand list (19)
ProjectYesterday ⭐Today ⭐ΔState
trimstray/the-book-of-secret-knowledge222k🆕First seen
ChromeDevTools/chrome-devtools-mcp40k🆕First seen
multica-ai/multica30k🆕First seen
Lum1104/Understand-Anything16k🆕First seen
teng-lin/notebooklm-py14k🆕First seen
alireza0/s-ui8.9k🆕First seen
colbymchenry/codegraph8.4k13k+4111Strengthening
multica-ai/andrej-karpathy-skills140k143k+2493Continuous
Imbad0202/academic-research-skills16k18k+2360Strengthening
dotnet/skills2.1k🆕First seen
obra/superpowers200k201k+1509Continuous
antoinezambelli/forge1.4k🆕First seen
rohitg00/ai-engineering-from-scratch9.2k10k+1288Strengthening
anthropics/claude-plugins-official21k22k+1070Strengthening
truelockmc/streambert2.6k3.7k+1055Strengthening
msitarzewski/agency-agents103k103k+912Continuous
rmyndharis/OpenWA4.6k5.3k+681Strengthening
HKUDS/CLI-Anything38k39k+616Continuous
can1357/oh-my-pi5.3k5.7k+466Strengthening

This week

New AI projects like `humanlayer/12-factor-agents` saw a major surge this week, while `colbymchenry/codegraph` demonstrates sustained interest by appearing on all timeframes.

5 Strengthening · 5 Continuous · 1 Stable · 3 First seen
Expand list (14)
ProjectYesterday ⭐Today ⭐ΔState
humanlayer/12-factor-agents22k🆕First seen
supertone-inc/supertonic9.1k🆕First seen
HKUDS/ViMax6.4k🆕First seen
colbymchenry/codegraph8.4k13k+4111Strengthening
Imbad0202/academic-research-skills16k18k+2360Strengthening
mattpocock/skills96k99k+2104Continuous
obra/superpowers200k201k+1509Continuous
tinyhumansai/openhuman23k25k+1310Strengthening
ruvnet/RuView62k63k+1228Continuous
rohitg00/agentmemory15k16k+779Strengthening
CloakHQ/CloakBrowser17k18k+682Strengthening
datawhalechina/easy-vibe13k14k+260Continuous
oven-sh/bun92k92k+62Stable
facebook/pyrefly6.3k6.4k+41Continuous

This month

The monthly trend reflects sustained accumulation in educational resources, such as `multica-ai/andrej-karpathy-skills`, and developer tools like `colbymchenry/codegraph`.

4 Strengthening · 11 Continuous · 1 Stable · 1 First seen
Expand list (17)
ProjectYesterday ⭐Today ⭐ΔState
huggingface/ml-intern9.7k🆕First seen
colbymchenry/codegraph8.4k13k+4111Strengthening
multica-ai/andrej-karpathy-skills140k143k+2493Continuous
Imbad0202/academic-research-skills16k18k+2360Strengthening
mattpocock/skills96k99k+2104Continuous
rohitg00/agentmemory15k16k+779Strengthening
CloakHQ/CloakBrowser17k18k+681Strengthening
TauricResearch/TradingAgents78k78k+486Continuous
Alishahryar1/free-claude-code27k27k+444Continuous
addyosmani/agent-skills44k44k+357Continuous
decolua/9router13k13k+343Continuous
ruvnet/ruflo53k54k+339Continuous
AIDC-AI/Pixelle-Video19k19k+297Continuous
anthropics/financial-services26k26k+258Continuous
Z4nzu/hackingtool76k76k+185Stable
Anil-matcha/Open-Generative-AI16k16k+184Continuous
soxoj/maigret30k30k+150Continuous

Categories· 5

AI Coding Tools & Agents(18)

This category is dominated by the formalization of an agent "skill" ecosystem, heavily centered on Anthropic's Claude. The focus has shifted from building monolithic agents to creating and managing shareable, componentized capabilities.

  • Anthropic's official directory for Claude plugins. Its creation is a major signal that Anthropic is formalizing its agent ecosystem, moving towards an "app store" model for agent capabilities, and consolidating its platform.

    new⭐ 22k
  • A tool to create a local knowledge graph of a codebase for AI agents. This is notable as a practical attempt to reduce token costs and improve context quality for agents like Claude by pre-processing information locally.

    rising⭐ 13k
  • A prompt file for Claude based on Andrej Karpathy's writings about LLM weaknesses. It's an example of the "skill" trend, where expert knowledge is distilled into a reusable prompt for agents.

    repeated⭐ 143k
  • A set of prompts for Claude defining a multi-step academic research workflow. Like the Karpathy skills repo, it exemplifies the pattern of packaging domain-specific processes as reusable agent skills.

    rising⭐ 18k
  • A repository from the .NET team for AI agent skills related to C#/.NET development. This shows a major language platform investing in the agent ecosystem, validating the "skill" as a primary shareable unit.

    new⭐ 2.1k
  • A library for providing persistent memory to AI agents, allowing them to maintain context across sessions. This project addresses a core architectural limitation of stateless LLMs and is a foundational component for more capable agents.

    repeated⭐ 16k
  • Uses WiFi signals for spatial awareness and presence detection without cameras. While novel, its connection to the core AI coding agent trend is tangential. It likely trended based on its unique application of signal processing.

    new⭐ 63k
  • Official tooling from the Chrome DevTools team to allow coding agents to interact with browser developer tools. A significant integration that gives agents programmatic access to a core part of the web development workflow.

    new⭐ 40k
  • An automated short video generation engine, primarily targeting the Chinese market. It represents the application of generative AI to content creation, a parallel and commercially significant trend to coding agents.

    cooling⭐ 19k
  • An unofficial Python client for Google's NotebookLM. It provides programmatic access to a knowledge-grounding tool, making it potentially useful for RAG or agent-based research tasks that require external context.

    new⭐ 14k
  • A command-line AI coding agent. It is another entry in a crowded space of terminal-based agents, differentiating with features like hash-anchored edits and LSP integration to improve code modification reliability.

    new⭐ 5.7k
  • An open-source platform for managing teams of AI agents. This signals a move up the stack from building individual agents to orchestrating multiple agents, a sign of the ecosystem's maturation towards 'AgentOps'.

    new⭐ 30k
  • An academic project demonstrating an agentic approach to video generation, with different agents for roles like director and screenwriter. This is a conceptual exploration of multi-agent systems for creative tasks.

    new⭐ 6.4k
  • A proxy server that routes AI requests to various free and paid model providers. This tool addresses the significant cost of using proprietary models by abstracting access and providing fallbacks to free alternatives.

    cooling⭐ 13k
  • A collection of prompts and configurations for AI coding agents by a well-known developer. Its popularity stems from the author's reputation and the high demand for production-ready, reusable agent skills.

    cooling⭐ 44k
  • An orchestration platform specifically for Claude-based agents. Its existence further solidifies Claude's growing ecosystem and highlights the emerging need for specialized tools to manage complex, multi-agent workflows.

    cooling⭐ 54k
  • A self-hosted web UI for various open-source image and video generation models. It's an aggregator and interface, appealing to users who want to avoid paid platforms and content filters.

    cooling⭐ 16k
  • A tool for generating explorable knowledge graphs from codebases. It taps into the need for better code comprehension tools that can be used interactively or fed to AI agents for improved context.

    new⭐ 16k

Prompt & Skill Libraries(2)

This small but highly concentrated category is the clearest signal of the day's main trend. Both `mattpocock/skills` and `obra/superpowers` represent the move to codify and share agent capabilities as reusable components.

  • A personal collection of prompts and configurations for Claude from a well-known TypeScript educator. It’s a prime example of the trend where individual experts share their agent "skillsets," validating the componentized approach.

    repeated⭐ 99k
  • A framework and methodology for developing agent "skills." It aims to bring more structure and engineering discipline to the process of creating agent capabilities, moving beyond simple, ad-hoc prompt collections.

    rising⭐ 201k

Other(18)

This is a diverse mix of perennial 'hacker' tools (`CloakHQ/CloakBrowser`, `soxoj/maigret`), high-level agent design principles (`12-factor-agents`), and noise (`anthropics/financial-services`). It highlights activity outside the core agent 'skill' trend.

  • A project aiming to create a "personal AI super intelligence". The description is aspirational and lacks technical specifics, suggesting it's more of a vision statement than a concrete tool at this stage. Its trend is likely speculative.

    new⭐ 25k
  • A modified Chromium browser designed to evade bot detection systems. It's a pragmatic tool for web scraping and automation, reflecting a persistent developer demand for tools that can bypass anti-automation measures.

    repeated⭐ 18k
  • An empty repository from Anthropic. Its name suggests a potential future focus area, but its trending status is likely due to speculation or a mistake, as it contains no code. This is currently noise.

    cooling⭐ 26k
  • A research project from HKUDS aiming to make any command-line tool usable by AI agents. This addresses a key challenge in agent capabilities: interfacing with the vast ecosystem of existing software.

    new⭐ 39k
  • A conceptual project from Hugging Face for an AI agent that can perform ML engineering tasks. It represents the ambition of major AI players to automate their own internal development and research workflows.

    cooling⭐ 9.7k
  • A self-hosted gateway for the WhatsApp API. This type of project is consistently popular, providing developers with programmatic access to closed messaging platforms for building bots and third-party integrations.

    new⭐ 5.3k
  • An educational repository for learning AI engineering. Its trend is likely driven by the author's existing following (from `agentmemory`) and the high demand for structured, practical learning materials on this topic.

    new⭐ 10k
  • Proposes a set of principles for building production-ready LLM applications, analogous to the classic "12-Factor App" methodology. This addresses the growing need for engineering discipline and best practices in the agent space.

    new⭐ 22k
  • A beginner-friendly coding course from a Chinese open-source community. Its trend is likely regional and reflects the community's efforts to promote modern coding practices to a new generation of developers.

    new⭐ 14k
  • An OSINT tool that finds user profiles across thousands of websites based on a single username. A classic example of the durable appeal of reconnaissance and digital forensics tools on GitHub.

    cooling⭐ 30k
  • A fast Python type checker and language server from Meta. Its trend is driven by Meta's backing and the ongoing need for high-performance developer tooling in the large and mature Python ecosystem.

    new⭐ 6.4k
  • A Python framework for agentic workflows and LLM tool-calling. It is one of many new frameworks trying to standardize how agents are built, competing in a rapidly crowding and competitive space.

    new⭐ 1.4k
  • A framework for using LLM-based agents for financial trading. This represents the application of agent technology to a high-stakes, specialized domain, attracting interest from both finance and AI communities.

    cooling⭐ 78k
  • An aggregator repository that collects various scripts and tools for penetration testing and hacking. These "all-in-one" collections are perennially popular but often consist of repackaged, pre-existing tools rather than new technology.

    cooling⭐ 76k
  • A curated list of cheatsheets, tools, and guides for developers and sysadmins. This is a classic "awesome list" style repository; its enduring value is in curation and community contributions, not original code.

    new⭐ 222k
  • An Electron-based desktop application for streaming media content, likely from unofficial sources. Its popularity is driven by user demand for free, aggregated access to media, bypassing official subscription services.

    new⭐ 3.7k
  • A collection of pre-defined agent personas for different business roles (e.g., frontend developer, community manager). This is another take on the "skill" concept, framed as complete agent personalities with specific processes.

    new⭐ 103k
  • A web panel for managing proxy tools like SagerNet and Sing-Box, which are popular for circumventing internet censorship. Its trend is likely driven by users in regions with restricted internet access.

    new⭐ 8.9k

On-device & Multimodal AI(2)

This category shows niche interest in cost-avoidance (`Alishahryar1/free-claude-code`) and offline capabilities (`supertone-inc/supertonic`). The signals here are weaker and overshadowed by the cloud-centric agent ecosystem trend.

  • A tool providing free command-line and VSCode access to a Claude model, likely by wrapping a free public endpoint. It highlights the persistent user desire to avoid API costs for powerful proprietary models.

    cooling⭐ 27k
  • A fast, on-device, multilingual text-to-speech engine running via ONNX. Its presence shows continued interest in high-performance, private, and offline AI capabilities, particularly for voice and accessibility applications.

    new⭐ 9.1k

AI Infra & Protocols(1)

With only one repository, `oven-sh/bun`, this category does not show a specific AI infrastructure trend today. Bun's popularity is part of the broader high-performance JavaScript tooling story.

  • A high-performance JavaScript runtime, bundler, and package manager. Its inclusion in AI trends is incidental; its popularity is part of a larger story about performance and consolidation in the JavaScript tooling ecosystem.

    new⭐ 92k

Recent reports

  • 2026-06-03

    generated 2026-06-03T17:05:52Z · rules v1 · repos 45 · signals 45

  • 2026-06-02

    generated 2026-06-02T16:49:13Z · rules v1 · repos 43 · signals 43

  • 2026-06-01

    generated 2026-06-01T17:57:29Z · rules v1 · repos 44 · signals 44

  • 2026-05-31

    generated 2026-05-31T14:03:33Z · rules v1 · repos 44 · signals 44

  • 2026-05-30

    generated 2026-05-30T13:57:40Z · rules v1 · repos 48 · signals 48

  • 2026-05-29

    generated 2026-05-29T16:02:16Z · rules v1 · repos 46 · signals 46

  • 2026-05-28

    generated 2026-05-28T16:19:16Z · rules v1 · repos 48 · signals 48

  • 2026-05-27

    generated 2026-05-27T16:04:14Z · rules v1 · repos 45 · signals 45

  • 2026-05-26

    generated 2026-05-26T16:01:07Z · rules v1 · repos 45 · signals 45

  • 2026-05-25

    generated 2026-05-25T15:24:58Z · rules v1 · repos 45 · signals 45

  • 2026-05-24

    generated 2026-05-24T13:51:52Z · rules v1 · repos 43 · signals 43

  • 2026-05-23

    generated 2026-05-23T13:55:04Z · rules v1 · repos 42 · signals 42

  • 2026-05-22

    generated 2026-05-22T14:56:21Z · rules v1 · repos 40 · signals 40

  • 2026-05-21 · current

    generated 2026-05-21T15:46:27Z · rules v1 · repos 41 · signals 41

  • 2026-05-20

    generated 2026-05-20T15:48:11Z · rules v1 · repos 36 · signals 36

  • 2026-05-19

    generated 2026-05-19T15:42:42Z · rules v1 · repos 37 · signals 37

  • 2026-05-18

    generated 2026-05-18T15:51:10Z · rules v1 · repos 39 · signals 39

  • 2026-05-17

    generated 2026-05-17T13:49:34Z · rules v1 · repos 45 · signals 45

  • 2026-05-16

    generated 2026-05-16T13:49:33Z · rules v1 · repos 34 · signals 34

  • 2026-05-15

    generated 2026-05-15T14:26:57Z · rules v1 · repos 38 · signals 38

  • 2026-05-14

    generated 2026-05-14T14:32:26Z · rules v1 · repos 42 · signals 42

  • 2026-05-13

    generated 2026-05-13T14:53:18Z · rules v1 · repos 44 · signals 44

  • 2026-05-12

    generated 2026-05-12T14:42:52Z · rules v1 · repos 36 · signals 36

  • 2026-05-11 · fallback

    generated 2026-05-11T15:30:34Z · rules v1 · repos 40 · signals 40

  • 2026-05-10

    generated 2026-05-10T13:46:39Z · rules v1 · repos 37 · signals 37

  • 2026-05-09

    generated 2026-05-09T13:43:42Z · rules v1 · repos 40 · signals 40

  • 2026-05-08

    generated 2026-05-08T14:08:16Z · rules v1 · repos 38 · signals 38

  • 2026-05-07

    generated 2026-05-07T14:38:09Z · rules v1 · repos 37 · signals 37

  • 2026-05-06

    generated 2026-05-06T14:34:27Z · rules v1 · repos 36 · signals 36

  • 2026-05-04

    generated 2026-05-04T14:22:31Z · rules v1 · repos 36 · signals 36

  • 2026-05-03

    generated 2026-05-03T13:40:38Z · rules v1 · repos 29 · signals 29

  • 2026-05-02 · fallback

    generated 2026-05-02T13:39:14Z · rules v1 · repos 30 · signals 30

  • 2026-05-01 · fallback

    generated 2026-05-01T13:46:05Z · rules v1 · repos 32 · signals 32

  • 2026-04-30

    generated 2026-04-30T14:19:05Z · rules v1 · repos 36 · signals 36

  • 2026-04-29

    generated 2026-04-29T14:23:30Z · rules v1 · repos 36 · signals 36

  • 2026-04-28

    generated 2026-04-28T14:34:52Z · rules v1 · repos 37 · signals 37

  • 2026-04-27

    generated 2026-04-27T14:20:15Z · rules v1 · repos 38 · signals 38

  • 2026-04-26

    generated 2026-04-26T13:24:10Z · rules v1 · repos 37 · signals 37

  • 2026-04-25

    generated 2026-04-25T13:22:57Z · rules v1 · repos 41 · signals 41

  • 2026-04-24

    generated 2026-04-24T13:58:04Z · rules v1 · repos 40 · signals 40

  • 2026-04-23

    generated 2026-04-23T14:04:35Z · rules v1 · repos 43 · signals 43

  • 2026-04-22

    generated 2026-04-22T14:02:49Z · rules v1 · repos 41 · signals 41

  • 2026-04-21

    generated 2026-04-22T07:58:11Z · rules v1 · repos 40 · signals 40

  • 2026-04-20

    generated 2026-04-21T09:23:28Z · rules v1 · repos 40 · signals 40

  • 2026-04-19

    generated 2026-04-19T08:51:11Z · rules v1 · repos 36 · signals 36

  • 2026-04-18

    generated 2026-04-18T08:45:02Z · rules v1 · repos 39 · signals 39

  • 2026-04-17

    generated 2026-04-17T09:14:47Z · rules v1 · repos 40 · signals 40

  • 2026-04-03

    generated 2026-05-13T03:24:00Z · rules v1 · repos 26 · signals 26

  • 2026-04-02 · degraded

    generated 2026-04-03T19:29:39Z · rules v1 · repos 26 · signals 26

  • 2026-04-01 · fallback

    generated 2026-04-03T19:29:39Z · rules v1 · repos 26 · signals 26