2026-05-23

Denoise

The AI agent ecosystem is standardizing around "skills" as the portable unit of capability, moving from simple prompts to structured, shareable toolsets.

Watch how platform owners like Anthropic and language maintainers like .NET are building official skill libraries, while developers focus on knowledge graphs for deeper code context.

2026-05-232026-05-23T13:55:04Zrules v1Healthyrepos 42signals 42

Top 3 changes

  • anthropics/claude-plugins-official shows major model providers are now building curated, App Store-like ecosystems to control quality and distribution.
  • Lum1104/Understand-Anything shows the shift in context augmentation from basic RAG to structured knowledge graphs for complex codebases.
  • supertone-inc/supertonic shows a parallel trend of pushing high-performance AI tasks like TTS onto local devices, contrasting with the cloud-centric agent model.

Strategic insights

#01The term 'skill' has become the dominant abstraction for extending AI agents, adopted by Anthropic, Microsoft (.NET), and independent developers like mattpocock. This signals convergence on a standard for packaging agent capabilities.
#02Model providers are becoming platform builders. Anthropic's `claude-plugins-official` and Google's `chrome-devtools-mcp` are moves to create walled gardens or curated ecosystems, shifting value from open models to controlled platforms.
#03Knowledge graphs are emerging as the preferred method for providing deep context to code agents. The rise of `Lum1104/Understand-Anything` and `colbymchenry/codegraph` indicates a move beyond simple file-based context retrieval.
#04The trend list contains significant noise. Perennial utilities like `yt-dlp` and long-standing curated lists like `the-book-of-secret-knowledge` are trending again, indicating a lack of novel breakthroughs and recycling of established tools.

Trend tracking

Today

New AI skill-compilation repositories are surging, led by the massive debut of `multica-ai/andrej-karpathy-skills` and related projects like `Anthropic-Cybersecurity-Skills`.

5 Strengthening · 3 Continuous · 2 Stable · 6 First seen
Expand list (16)
ProjectYesterday ⭐Today ⭐ΔState
multica-ai/andrej-karpathy-skills149k🆕First seen
multica-ai/multica32k🆕First seen
mukul975/Anthropic-Cybersecurity-Skills7.0k🆕First seen
presenton/presenton6.1k🆕First seen
janestreet/magic-trace5.7k🆕First seen
colbymchenry/codegraph16k18k+2336Strengthening
Lum1104/Understand-Anything18k20k+2270Strengthening
anthropics/claude-plugins-official24k26k+1914Strengthening
NVlabs/LongLive1.7k🆕First seen
rohitg00/ai-engineering-from-scratch11k13k+1543Strengthening
yt-dlp/yt-dlp164k165k+682Stable
trimstray/the-book-of-secret-knowledge223k224k+577Stable
Fincept-Corporation/FinceptTerminal22k23k+497Continuous
ChromeDevTools/chrome-devtools-mcp41k41k+396Continuous
odoo/odoo51k51k+351Continuous
dotnet/skills2.4k2.7k+242Strengthening

This week

AI agent and skill-focused repositories show persistent weekly strength, with new entries `K-Dense-AI/scientific-agent-skills` and `Lum1104/Understand-Anything` establishing momentum.

6 Strengthening · 6 Continuous · 3 First seen
Expand list (15)
ProjectYesterday ⭐Today ⭐ΔState
K-Dense-AI/scientific-agent-skills25k🆕First seen
Lum1104/Understand-Anything20k🆕First seen
can1357/oh-my-pi6.6k🆕First seen
colbymchenry/codegraph16k18k+2336Strengthening
rohitg00/ai-engineering-from-scratch11k13k+1544Strengthening
obra/superpowers203k204k+1044Continuous
ruvnet/RuView64k64k+822Continuous
CloakHQ/CloakBrowser19k19k+662Strengthening
Imbad0202/academic-research-skills19k19k+580Strengthening
tinyhumansai/openhuman25k26k+579Continuous
rohitg00/agentmemory16k17k+378Continuous
supertone-inc/supertonic9.3k9.7k+322Strengthening
HKUDS/ViMax6.6k6.8k+207Strengthening
datawhalechina/easy-vibe14k14k+206Continuous
humanlayer/12-factor-agents22k22k+142Continuous

This month

The AI "skills" compilation trend is entrenching itself, seen with new list `ComposioHQ/awesome-codex-skills`, while `colbymchenry/codegraph` shows sustained growth across all timeframes.

3 Strengthening · 9 Continuous · 5 Stable · 3 First seen
Expand list (20)
ProjectYesterday ⭐Today ⭐ΔState
ComposioHQ/awesome-codex-skills11k🆕First seen
HKUDS/ViMax6.8k🆕First seen
mattpocock/sandcastle4.8k🆕First seen
multica-ai/andrej-karpathy-skills145k149k+3158Continuous
colbymchenry/codegraph16k18k+2336Strengthening
mattpocock/skills101k102k+1376Continuous
CloakHQ/CloakBrowser19k19k+662Strengthening
Imbad0202/academic-research-skills19k19k+580Strengthening
Alishahryar1/free-claude-code28k28k+522Continuous
rohitg00/agentmemory16k17k+378Continuous
TauricResearch/TradingAgents78k79k+280Stable
decolua/9router13k14k+265Continuous
anthropics/financial-services27k27k+255Continuous
addyosmani/agent-skills45k45k+243Continuous
ruvnet/ruflo54k54k+229Stable
AIDC-AI/Pixelle-Video19k19k+155Continuous
Z4nzu/hackingtool76k76k+146Stable
refactoringhq/tolaria11k11k+104Continuous
soxoj/maigret30k30k+101Stable
bytedance/UI-TARS-desktop35k35k+70Stable

Categories· 5

AI Coding Tools & Agents(21)

This category is dominated by ecosystem-building rather than new model creation. The focus is on tooling to make agents useful: knowledge graphs for code understanding (`Understand-Anything`, `codegraph`), official skill/plugin directories (`claude-plugins-official`), and standardized skill formats (`dotnet/skills`).

  • Transforms codebases into interactive knowledge graphs for AI agents. This approach to providing context is more structured than simple file retrieval, representing a move toward deeper, semantic code understanding. Its rise signals a key direction for improving agent capabilities.

    rising⭐ 20k
  • Similar to Understand-Anything, this project creates a code knowledge graph. It emphasizes pre-indexing and local operation to reduce token usage and improve performance for various agents. Trending alongside a similar tool confirms the pattern of structured context becoming critical.

    rising⭐ 18k
  • This is Anthropic's official, managed directory for Claude Code plugins. It's a significant platform play, aiming to create a curated, high-quality ecosystem similar to an app store, which could centralize and control the distribution of agent capabilities.

    new⭐ 26k
  • A structured collection of prompts and workflows for academic research, packaged as 'skills' for Claude Code. It exemplifies the trend of codifying domain expertise into reusable components for AI agents, though it is a content library rather than a tool.

    repeated⭐ 19k
  • This project uses commodity WiFi signals for spatial intelligence and presence detection without cameras. It's an interesting application of signal processing, but its presence on a list dominated by software-based AI agents is an anomaly.

    new⭐ 64k
  • Provides persistent memory for AI coding agents, a critical component for building stateful, multi-step autonomous systems. Its repeated appearance on trending lists indicates sustained developer interest in solving the core challenge of agent continuity.

    repeated⭐ 17k
  • A single markdown file containing prompts and instructions for Claude Code, based on Andrej Karpathy's analysis of LLM pitfalls. It highlights the demand for expert-curated 'skills' to improve agent reliability and behavior.

    repeated⭐ 149k
  • An official Microsoft repository for .NET and C# skills for AI coding agents. This is a strong signal that major programming ecosystems are formally embracing the 'skill' abstraction and building first-party integrations for AI agents.

    new⭐ 2.7k
  • A curated list of skills for the Codex agent. While useful, the 'awesome-list' format is a common pattern in maturing technology areas and is less indicative of innovation than a new framework or tool.

    cooling⭐ 11k
  • This repository provides Chrome DevTools specifically for coding agents. It's an initiative by a major platform player (Google) to provide better integration points for agents, signaling the importance of browser automation in agent workflows.

    new⭐ 41k
  • An agentic video generation framework that divides the task among specialized agents (Director, Screenwriter, etc.). Its repeated trend shows continued interest in structured, multi-agent approaches for complex creative tasks beyond code generation.

    repeated⭐ 6.8k
  • A large, structured collection of cybersecurity skills for agents, mapped to industry frameworks like MITRE ATT&CK. This demonstrates the application of the 'skill' concept to a highly specialized, compliance-driven domain.

    new⭐ 7.0k
  • An open-source platform for managing coding agents as if they were team members. It aims to solve the orchestration problem, moving beyond single-agent scripts to coordinated, task-driven workflows. Addresses a key operational challenge for agent deployment.

    new⭐ 32k
  • An automated short video generation engine. Like ViMax, it represents interest in applying agent-based systems to content creation, a popular but crowded space. Its value depends on the quality of its output compared to established alternatives.

    cooling⭐ 19k
  • A tool to route API calls for AI coding assistants to free or cheaper model providers. This is a cost-arbitrage utility, reflecting user sensitivity to API pricing but not contributing new capabilities to the agent ecosystem itself.

    cooling⭐ 14k
  • An infrastructure project from NVIDIA Labs for long video generation. Coming from a major research lab, it's worth monitoring, but as an infrastructure piece, its immediate utility for general developers may be limited.

    new⭐ 1.7k
  • A command-line AI coding agent with features like LSP integration and sub-agent support. It represents a more developer-centric, terminal-first approach to building agents, appealing to users who prefer CLI workflows.

    new⭐ 6.6k
  • A well-known and long-maintained command-line video downloader. Its appearance on the trending list is likely due to external events or mentions rather than any fundamental innovation within the project itself. This is background noise.

    new⭐ 165k
  • An agent orchestration platform, similar in goal to multica. The presence of multiple orchestration tools indicates that managing and coordinating agents is becoming a significant problem as developers move from experiments to production systems.

    cooling⭐ 54k
  • A collection of engineering-focused skills for coding agents from a well-known developer. Like other skill libraries, it contributes to the ecosystem, but its value is in the curated content rather than a novel technical approach.

    cooling⭐ 45k
  • A TypeScript library for orchestrating sandboxed coding agents. It focuses on safety and isolation, a critical concern for autonomous agents that execute code. This addresses a practical security need in the growing agent ecosystem.

    cooling⭐ 4.8k

Other(14)

A diverse group showing niche AI applications and developer tooling. `CloakHQ/CloakBrowser` tackles the persistent bot detection problem, while `humanlayer/12-factor-agents` attempts to establish best practices for a maturing field. The presence of older, well-known repos like `odoo` and `the-book-of-secret-knowledge` points to recycled interest rather than new trends.

  • Presents itself as a 'Personal AI super intelligence' focused on privacy. The claims are broad and need verification. This type of project often attracts initial interest based on its ambitious premise rather than demonstrated capability.

    new⭐ 26k
  • An empty repository from Anthropic. Its trending status is likely due to speculation or a mistake. It holds no functional value at present but indicates interest in Anthropic's moves in the finance vertical.

    cooling⭐ 27k
  • An educational repository for learning AI engineering concepts. Its popularity reflects a strong demand for foundational knowledge as more developers enter the AI space. It's a learning resource, not a production tool.

    rising⭐ 13k
  • A stealth-focused Chromium browser designed to defeat bot detection systems. It operates at the source code level to modify fingerprints. This is a specialized tool in the continuous arms race between automation and detection.

    repeated⭐ 19k
  • A financial analysis application. Its presence, along with other finance-related repos, suggests that finance is a key area where developers are applying modern data analysis and AI techniques.

    new⭐ 23k
  • An attempt to define a set of principles for building production-ready LLM applications, analogous to the classic '12-Factor App'. This signals that the field is maturing and looking for established engineering best practices.

    new⭐ 22k
  • An open-source AI presentation generator. This is an application of generative AI to a common business task, competing in a space with established commercial players. Its value lies in its open-source nature.

    new⭐ 6.1k
  • An OSINT tool that collects information on individuals by searching usernames across thousands of websites. This is a long-standing tool in the cybersecurity and privacy community, not a new development.

    cooling⭐ 30k
  • A coding course for beginners. As an educational resource, its trend is likely driven by a new cohort of students or a promotional event in its target community, rather than a technical innovation.

    new⭐ 14k
  • A framework for building multi-agent systems for financial trading. This combines two major trends: agentic AI and algorithmic trading. It represents a specialized, high-stakes application of agent technology.

    cooling⭐ 79k
  • A very large, long-standing curated list of tools, cheatsheets, and hacks for developers and administrators. Its periodic trending is a sign of evergreen content discovery, not a new development.

    new⭐ 224k
  • A popular open-source suite of business management apps (ERP, CRM). Its appearance on GitHub trending is unusual for a mature project and likely tied to a major new release or conference.

    new⭐ 51k
  • A high-resolution tracing tool from Jane Street for analyzing process behavior. This is a sophisticated, low-level performance analysis tool, appealing to developers working on high-performance systems. Its relevance is to systems engineering, not the mainstream AI trend.

    new⭐ 5.7k
  • An all-in-one collection of hacking tools. These types of repositories, which bundle existing tools, are common but often of mixed quality. They cater to beginners in cybersecurity but rarely offer novel capabilities.

    cooling⭐ 76k

Prompt & Skill Libraries(3)

This category highlights the community's effort to standardize agent capabilities. Rather than just lists of prompts, projects like `mattpocock/skills` and `obra/superpowers` provide structured frameworks, treating 'skills' as first-class citizens in agent development.

  • A collection of practical skills for agents, curated by a prominent developer. Its popularity highlights the demand for high-quality, trusted prompt engineering, and reinforces the 'skill' as the key unit of agent enhancement.

    cooling⭐ 102k
  • This project proposes a framework and methodology for developing with 'agentic skills'. It attempts to formalize the process of building agent capabilities, moving beyond ad-hoc prompt collections to a more systematic software development lifecycle.

    new⭐ 204k
  • A set of agent skills specifically for scientific and financial tasks. This follows the pattern of creating domain-specific skill libraries, suggesting the agent ecosystem is fragmenting into specialized verticals.

    new⭐ 25k

On-device & Multimodal AI(3)

This small set highlights the push for local AI execution. `supertone-inc/supertonic` provides on-device TTS via ONNX, a tangible example of moving complex models to the edge. This contrasts with wrappers like `free-claude-code`, which simply provide alternative access to cloud models.

  • An on-device, multilingual text-to-speech engine using the ONNX runtime. This is a practical implementation of running sophisticated AI models locally, offering speed and privacy benefits over cloud-based APIs. It represents a significant trend in edge AI.

    new⭐ 9.7k
  • A wrapper that provides free access to the claude-code model through various interfaces like a CLI and VSCode extension. This is a utility for circumventing API costs, not a foundational technology.

    cooling⭐ 28k
  • An open-source stack for multimodal AI agents from ByteDance. As a contribution from a major tech company, it provides an integrated infrastructure for building agents that can understand and interact with user interfaces.

    cooling⭐ 35k

Memory, RAG & Context(1)

The single entry, `refactoringhq/tolaria`, is a desktop application for managing local knowledge bases. Its presence suggests a demand for private, user-controlled knowledge management, running parallel to the developer-focused, programmatic memory solutions like `rohitg00/agentmemory` seen in other categories.

  • A desktop application for managing markdown-based knowledge bases. It caters to the need for local, private knowledge management, a user-centric parallel to the programmatic agent memory solutions seen elsewhere.

    cooling⭐ 11k

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 · current

    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

    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