Skip to content

Skills Catalog

Welcome to the Hermes Skills Catalog - your directory of community-contributed, validated skills that extend what Hermes can do. Skills encode repeatable expertise into shareable packages that anyone in the Hermes community can install and use.

Skill Quality Tiers

Not all skills are created equal. We use three quality tiers to help you understand what you're installing:

Production Tier 🟢

Production-tier skills meet all of these criteria:

  • Tested by at least three independent community members
  • Includes comprehensive error handling for all known failure modes
  • Documentation covers setup, invocation, expected output, and troubleshooting
  • Confirmation gates on all write/destructive operations
  • Maintained actively (updated within 30 days of reported issues)
  • Pinned dependencies and explicit version requirements
  • Audit trail for all data access

These skills are safe for production use. You can rely on them as components of your automated workflows. Look for the 🟢 indicator in the catalog.

Beta Tier 🟡

Beta-tier skills are functional and well-tested by their authors but haven't completed community validation:

  • Tested by the author and at least one other person
  • Handles common error cases but may have gaps in edge-case handling
  • Documentation covers basic usage but may lack troubleshooting depth
  • May lack comprehensive confirmation gates
  • Updated within 90 days

These skills are suitable for supervised use. They'll save you time but keep an eye on them - especially in the first few runs. Look for the 🟡 indicator.

Community Tier 🔵

Community-tier skills are shared in good faith but haven't completed formal validation:

  • Published by a community member
  • May have been tested only on the author's setup
  • Documentation may be minimal
  • Error handling may be incomplete
  • Best suited for learning, inspiration, and adaptation - not production reliance

These skills are valuable for the community but require due diligence before relying on them. Look for the 🔵 indicator.

| RunComfy Agent Skills | 🟡 Beta | 30 | 61.1K | AI video, image-to-video, avatar video, video editing, music generation | prime-skills/runcomfy-agent-skills |

How to Evaluate a Skill Before Installing

Before you trust a skill with your data and credentials, do this five-minute review:

1. Read the Skill Description

Does it clearly state what it does, what connectors it needs, and what output it produces? A skill that can't explain itself in three sentences is a red flag.

2. Check the Required Permissions

What connectors and tools does the skill call? Does "Generate weekly report" really need write access to your CRM? If the permission scope exceeds the stated purpose, investigate before installing.

3. Review the Error Handling

Open the skill file and look for error handling blocks. Does it handle timeouts? Rate limits? Missing data? A skill with no error handling will fail silently or confusingly.

4. Look for Confirmation Gates

If the skill can modify data, does it require confirmation before doing so? Any skill that writes without explicit user approval is a production risk.

5. Check Recency and Maintenance

When was the skill last updated? A skill that hasn't been touched in 12 months may have broken dependencies or incompatible API versions. Check the changelog.

6. Read Community Feedback

Look for comments, issues, or reviews from other community members. "Works great on my setup" from one person is positive. "Breaks when the dataset exceeds 100 records" is actionable.

7. Test in a Safe Environment

Run the skill first in a read-only mode or with limited data. Before you let it loose on your production CRM, test it on your sandbox.

Curated Starter Pack

New to Hermes? These 10 skills are the most commonly recommended starting points. They cover essential workflows and are all Production-tier verified.

1. Daily Briefing

What it does: Your morning dashboard - calendar, priority emails, task list, and key metrics from connected services. One invocation replaces five separate checks.
Needs: Calendar, email, and optionally CRM or project management connectors.
Why start here: Replaces your morning routine with a single command. Immediate time savings.

2. Email Digest

What it does: Summarizes unread emails from a configurable time window, groups by thread, flags urgent items, and drafts suggested replies for routine messages.
Needs: Email connector (Gmail or Outlook).
Why start here: The average knowledge worker spends 2+ hours on email daily. This cuts it substantially.

3. Meeting Prep

What it does: Before a meeting, gathers relevant emails, documents, previous meeting notes, and action items related to the attendees and topic. Produces a one-page briefing.
Needs: Calendar, email, and optionally drive or project management connectors.
Why start here: Never walk into a meeting underprepared again.

4. Weekly Report Generator

What it does: Aggregates activity from your connected services into a structured weekly summary - tasks completed, meetings attended, key communications, metrics trends.
Needs: Project management, calendar, email (configurable).
Why start here: Automates the Friday afternoon ritual. Customize the template to match your team's format.

5. CRM Health Check

What it does: Pipeline analysis, stale deals, overdue follow-ups, and contact engagement scoring. Flags accounts that need attention.
Needs: CRM connector (HubSpot, Salesforce, or similar).
Why start here: Your CRM has data but Hermes turns it into action items.

6. Data to Chart

What it does: Takes a dataset (from any connector or upload) and generates appropriate visualizations - time series, distribution, comparison, correlation.
Needs: Any data source (accepts structured data from other skills).
Why start here: Visual insight without spreadsheet wrestling.

7. Content Drafting Assistant

What it does: Drafts blog posts, social media updates, newsletters, and other content following your brand voice. Includes tone calibration and audience targeting.
Needs: No specific connectors (optionally integrates with CMS or social platforms).
Why start here: Beat the blank page. All output is draft - you remain the editor.

8. Code Review Companion

What it does: Reviews code changes for bugs, security issues, style consistency, and documentation completeness. Provides actionable, specific feedback.
Needs: File access or repository connector.
Why start here: A second pair of eyes on every PR, catching issues before humans spend time reviewing.

9. Travel Planner

What it does: Given destination and dates, gathers flight options, hotel availability, weather forecasts, and local information. Organizes into a comparison view.
Needs: Calendar (for availability context).
Why start here: Research acceleration - you still book, but Hermes does the hunting.

10. Knowledge Base Q&A

What it does: Searches your organization's documentation, wikis, and shared drives to answer questions. "What's our vacation policy?" "How do I set up the VPN?"
Needs: Drive, wiki, or documentation connectors.
Why start here: Reduces the "just ask Bob" tax on your organization's experts.

Installing a Skill

Skills from the catalog install with a single command. From the catalog page, copy the install command and run it in your Hermes terminal. The skill and its dependencies are installed to your profile.

hermes skills install skill-id

After installation, configure any required connectors. Most skills include a setup guide that walks through connector authentication. Test with a dry run before relying on the skill in production workflows.

Contributing to the Catalog

The skills catalog thrives on community contributions. If you've built something useful:

  1. Follow the skill development best practices
  2. Test thoroughly in your environment
  3. Document setup, invocation, and expected output
  4. Remove environment-specific values (use placeholders)
  5. Submit through the catalog contribution process

Skills are reviewed by community maintainers before publication. The review checks for documentation completeness, error handling, security considerations, and community value - not whether every edge case is handled (that's what quality tiers communicate).

Finding More Skills

The catalog here represents community-validated skills. Additional skills are discoverable through:

  • skills.sh - Large open marketplace
  • agentskills.io - Curated premium and community skills
  • hermeshub - Official Hermes skill registry
  • skilldock.io - Enterprise-focused skill marketplace
  • GitHub - Search for "hermes-skill" or "hermes-skill-" prefixed repositories

See Skill Marketplaces for detailed guidance on each marketplace.

The catalog is a living resource. Skills are added weekly. Check back often, and consider contributing what you build.


This Hermes repo is one of the largest structured collections of public AI, automation, business, and technology documentation. Content remains attributed to original authors and repositories. Indexed and organized by www.CorpusIQ.io.

All Skills Catalog Pages

New - September 30, 2026 Sweep (8 publisher guides)

New - September 29, 2026 Sweep (64 publisher guides)

Previously unlinked pages (added 2026-10-06)

Linked from this index so the pages are reachable by crawl.

Skill setup guides

New - October 26, 2026 Sweep (14 publisher guides)

New - October 26, 2026 Evening Sweep (8 publisher guides)

New - November 1, 2026 Morning Sweep (14 publisher guides)

Fourteen repository clusters drained from the standing queue, documented for 179,773 combined installs across 476 indexed listings:

  • Syncfusion Angular UI Skills - Vendor Component Pack - 17,636 combined installs across 61 listings; 0 stars; no LICENSE file; Syncfusion's Angular component skills: one skill per component family for grid, scheduler, Gantt, charts and editors, 65 SKILL.md directories against 61 listings, with the commercial component licence separate from the skill files
  • ChatGPT Skills - 12-Suite Deterministic Toolkit - 16,721 combined installs across 106 listings; 126 stars; no LICENSE file; twelve shipped suites for OCR document processing, spreadsheets, file conversion and research work behind 106 indexed listing names, so most rows name capability with nothing to install
  • Huashu Skills - 22-Skill Content Production Pack - 16,023 combined installs across 37 listings; 2147 stars; MIT; 22 Chinese-first content production skills for HTML-first design, markdown to PDF, social cards and video scripting, against 37 listings that include 15 older names for skills already covered
  • VectorBT Backtesting Skills - Quant Workflow Setup - 15,152 combined installs across 6 listings; 217 stars; no LICENSE file; six quant backtesting skills over vectorbt, from data loading and indicator setup through portfolio construction and walk-forward evaluation, one directory per listing
  • Slack Skills Plugin - MCP Plus Nine Slack Skills - 15,188 combined installs across 9 listings; 163 stars; MIT; official Slack package: a hosted Slack MCP server plus nine skills split between using Slack and building on it, with Snyk W012 on the two skills that fetch live docs at task time
  • QwenCloud AI Skills - Multimodal Suite Setup - 15,120 combined installs across 10 listings; 44 stars; Apache-2.0; Qwen Cloud multimodal skills for vision, OCR, speech and image generation through the cloud API, ten directories behind ten listings
  • Full Stack Skills - 788-Skill Catalog for Agents - 14,932 combined installs across 102 listings; 718 stars; Apache-2.0; a navigation hub for 788 skill entries across 51 separately installable packages since the June 2026 split, so this repository is the directory rather than the install source
  • Supabase Pentest Skills - 24-Skill Audit Suite - 13,050 combined installs across 25 listings; 73 stars; NOASSERTION; 24 Supabase security-audit skills led by RLS and auth checks, where the README read-only claim is qualified because one skill creates a test user and the evidence directory holds sampled rows
  • Startup OS Skills - Sales, Product and Video Pack - 12,860 combined installs across 53 listings; 53 stars; no LICENSE file; 54 startup operating skills across sales, product, marketing and video, one directory per listing plus one unindexed directory
  • AgentMail Skills - Agent Email Inbox Setup - 10,705 combined installs across 11 listings; 25 stars; no LICENSE file; AgentMail email skills for inbox provisioning, send and receive, MCP configuration and framework adapters, with two deprecated alias listings inflating the total and Snyk W011 on the two package-install skills
  • Awesome Android Agent Skills - Compose and Kotlin Setup - 9,948 combined installs across 18 listings; 998 stars; Apache-2.0; 17 Android skills under .github/skills covering Compose, coroutines, ViewModel, architecture, migration, testing and accessibility, plus an Agent.md that every agent reads implicitly
  • Shenzjd Skills - DevOps and AI Image Pack - 7,887 combined installs across 16 listings; 4 stars; MIT; eight DevOps and AI image directories behind 16 listings, with a Snyk FAIL (E005) on the two skills that write executable workflow and hook files
  • Robust Skills - Architecture and Platform Suite - 7,780 combined installs across 12 listings; 61 stars; MIT; architecture and platform skills for DDD, hexagonal boundaries, contract-first APIs and state ownership, eleven directories plus one index-only listing
  • Awesome Finance Skills - Market Analysis Pack - 6,771 combined installs across 10 listings; 3094 stars; Apache-2.0; market analysis skills for news sentiment, alpha research and factor work, with Snyk W011 and W012 on the two skills that reach the network for live market data

New - October 31, 2026 Evening Sweep (16 publisher guides)

Sixteen repository clusters drained from the same queue the morning fire worked, documented for 332,009 combined installs across 515 indexed listings:

  • Awesome Codex Skills - Curated Codex Index - 24,951 combined installs across 61 listings; 17,150 stars; no LICENSE file; 880 in-tree SKILL.md directories behind 61 listings, a curated index drawn from many publishers with a Snyk Warn on three sampled entries
  • CXUU Golang Skills - Go Engineering Suite - 24,917 combined installs across 20 listings; 188 stars; Apache-2.0; Go skills covering concurrency, interfaces, testing and module layout, one directory per listing, with a Trust Hub Warn on the sampled go-interfaces skill
  • 9arm Skills - Thai Dev Agent Suite - 24,419 combined installs across 6 listings; 3,244 stars; no LICENSE file; six skills from a single practitioner portfolio led by scrutinize at 5,233 installs, one directory per listing
  • Neo4j Skills - Graph Database Agent Suite - 24,115 combined installs across 32 listings; 119 stars; MIT; thirty Neo4j skills across Cypher, modelling, imports, GraphRAG, five drivers, GDS, Aura, Kafka and Spark, plugin manifests for four harnesses, two listings with no directory, and a licence that reads MIT in the file but Apache-2.0 in the manifests
  • CC DevOps Skills - Kubernetes and IaC Suite - 22,236 combined installs across 31 listings; 330 stars; Apache-2.0; 31 DevOps skills across Kubernetes, Terraform, CI and log tooling, where ansible-validator warns on Gen Agent Trust Hub, Socket and Snyk
  • Aaronontheweb Dotnet Skills - Akka.NET Suite - 21,954 combined installs across 40 listings; 1,244 stars; MIT; Akka.NET and .NET engineering skills against 39 directories, a separate repository from the official dotnet/skills suite already in the catalog, with one Snyk Warn in the sample
  • Gemini CLI Skills - Repository Workspace Pack - 21,729 combined installs across 19 listings; 108,637 stars; Apache-2.0; Google's own CLI monorepo where the listings sit under .gemini/skills and the packages builtins rather than a top-level skills tree, with two Snyk Warns in the sample
  • ast-grep Skill - Structural Code Search - 20,794 combined installs across 2 listings; 900 stars; no LICENSE file; two skills only, ast-grep at 18,303 and outline at 2,491, carrying 20,794 installs between them
  • Metabase Skills - BI and Dashboard Agent Suite - 20,284 combined installs across 28 listings; 50,069 stars; NOASSERTION; skills split across resources/mcp/v2/skills and .claude/skills inside a 50,069-star analytics monorepo, five data-app listings with no directory, one Snyk Warn
  • Academic Writing Skills - Paper Workflow - 18,655 combined installs across 11 listings; 528 stars; no LICENSE file; academic writing skills led by latex-paper-en at 5,973 installs, with one listing that has no directory
  • Modular Skills - Mojo and MAX Agent Suite - 18,312 combined installs across 12 listings; 207 stars; NOASSERTION; Mojo and MAX skills in eleven directories, where the GitHub licence field reads NOASSERTION over an Apache 2.0 file with LLVM exceptions
  • Guard Skills - Code Safety Review Pack - 18,122 combined installs across 5 listings; 1,301 stars; MIT; five guardrail review skills led by clean-code-guard at 4,746 installs, five listings against five directories and every sampled verdict a Pass
  • Catalyst by Zoho Skills - Serverless Suite - 17,987 combined installs across 19 listings; 70 stars; Apache-2.0; Zoho Catalyst serverless skills covering functions, data store, auth and deployment, one directory per listing, with Warns on the 36-install tail listing
  • ClaudeKit Skills - Full Stack Agent Toolkit - 17,978 combined installs across 33 listings; 2,230 stars; no LICENSE file; 45 SKILL.md directories under .claude/skills covering document work, UI styling, MCP building and code review, with one listing and one directory that do not pair
  • Awesome Ecommerce Skills - Commerce Index - 17,780 combined installs across 178 listings; 60 stars; MIT; a flat install curve against 181 in-tree directories, where the top 20 sit at or above 144 installs and the leader reads 301
  • Layers Skills - Design System Agent Pack - 17,776 combined installs across 18 listings; 314 stars; MIT; nine design-system directories behind 18 listings, with nine bare-name listings at 1 to 2 installs that have no directory

New - October 31, 2026 Morning Sweep (16 publisher guides)

Sixteen repository clusters from the queue the October 30 evening probe left behind, documented for 536,147 combined installs across 835 indexed listings:

  • Jaganpro SF Skills - Salesforce Agent Suite - 43,852 combined installs across 46 listings; 424 stars; MIT; Apex, LWC, Flow, SOQL, metadata, deploy and Agentforce skills scored against fixed point budgets, where the repository is archived with a documented successor at forcedotcom/afv-library and nine indexed names have no directory
  • Anti-Slop - AI Copywriting Quality Skill - 41,307 combined installs across 6 listings; 7,670 stars; MIT; Six skills that strip generic AI tells from UI, copy, mobile layout and code comments, with a contrast checker script, a Hermes-aware installer and a Trust Hub Warn on the core skill the others reference
  • Browserbase Skills - Agent Browser Automation - 40,127 combined installs across 20 listings; 3,747 stars; no LICENSE file; the official Browserbase suite: sessions, fetch, search, tracing, UI testing and agent functions, where one sampled skill fails Gen Agent Trust Hub
  • JEO Skills - Presentation and Document Suite - 37,915 combined installs across 112 listings; 53 stars; no LICENSE file; a 112-listing flat .agent-skills tree weighted toward presentations and office deliverables, where rtk fails Gen Agent Trust Hub and the guide leads with it
  • CCFA Skills - Academic Paper and Research Suite - 33,935 combined installs across 24 listings; 3,358 stars; MIT; academic paper work across 24 listings: literature search, LaTeX templates for 141 venues, reviewer simulation and rebuttal, with six indexed names that have no directory
  • TanStack Agent Skills - Query and Router Setup - 32,849 combined installs across 4 listings; 228 stars; MIT; TanStack Query, Router, Start and integration conventions as four skills, each with a rules folder of numbered documents, all sampled verdicts clean
  • Chris Banes Skills - Compose and Kotlin Suite - 32,776 combined installs across 35 listings; 1,133 stars; Apache-2.0; Compose, coroutines, Gradle and Kotlin Multiplatform skills from a Compose maintainer, 20 directories against 35 indexed names with a README migration table covering 13 of the gaps
  • Paul Berg Agent Skills - Solidity and Foundry - 32,742 combined installs across 97 listings; 104 stars; MIT; Solidity and Foundry engineering skills plus code polish, commit and dependency-bump tooling over 43 directories against 97 indexed listings
  • Maggy Skills - Android and Kotlin Agent Suite - 32,037 combined installs across 73 listings; 710 stars; MIT; 73 Android and Kotlin skills with 25 slash commands, 13 hooks, 9 conditional rules and a Cortex MCP server, every indexed listing present in the tree
  • SAP Skills - BTP, CAP and ABAP Agent Suite - 31,778 combined installs across 45 listings; 482 stars; GPL-3.0; SAP BTP, CAP, ABAP and integration skills, 40 of them shipped as plugins with their own commands, agents and marketplace manifest
  • Tencent CloudBase Agent Skills - WeChat Dev - 30,797 combined installs across 37 listings; 80 stars; MIT; WeChat mini-program and CloudBase skills from Tencent, a separate repository from the cloudbase-skills guide already in the catalog, with one Snyk Fail inside the sample
  • OrchestKit - Orchestration and DevOps Skills - 30,386 combined installs across 151 listings; 300 stars; MIT; 115 orchestration and delivery skills with 161 hooks, 36 agents and a build mirror of the same skill set, aimed at plan, deploy, review and release gates
  • Cypress AI Toolkit - E2E Testing Skills - 30,426 combined installs across 5 listings; 46 stars; MIT; the official Cypress end-to-end testing toolkit as five skills plus a streamable-http MCP endpoint, with Snyk Warns on the authoring and cloud skills
  • Asyraf Hussin Skills - Clean Code Suite - 29,010 combined installs across 24 listings; 95 stars; MIT; 24 engineering-standards skills, each with a rules folder and metadata, carrying no scripts and no hooks, with one Socket Warn in the sample
  • Tigris Data Skills - Object Storage for Agents - 28,128 combined installs across 19 listings; 2 stars; MIT; Tigris object storage as 19 skills, where all 18 directories are paired with a .zip distribution package and the sample carries one Socket Warn
  • Xiaohongshu Skills - RED Content Operations - 28,082 combined installs across 137 listings; 515 stars; no LICENSE file; 137 RED content-operation skills in nine Chinese-named category folders, with three disagreeing counts between the README, the plugin manifest and the tree

New - October 30, 2026 Evening Sweep (16 publisher guides)

Sixteen publisher clusters from a 68-query topic probe, the largest install weight this catalog has recorded for a single fire: 1,101,389 combined installs across 1,717 indexed listings.

  • Text to CAD Skills - 3D Model and URDF Setup - 124,096 combined installs across 18 listings; 24,377 stars; MIT; text-to-CAD and URDF generation with a viewer and a robot-motion skill, six indexed listings with no directory at main
  • Context Engineering Kit - Prompt and MCP Setup - 110,374 combined installs across 138 listings; 1,795 stars; GPL-3.0; prompt engineering, context engineering and agent workflow patterns across 204 SKILL.md files on a master branch
  • Science Skills - PubMed, arXiv and Lab Data Setup - 87,871 combined installs across 42 listings; 3,363 stars; Apache-2.0; Google DeepMind science skills for literature search, genomics, structural biology and chemistry, every retrieval step delegated to bundled uv scripts
  • SecondSky Claude Skills - Backend and API Suite Setup - 85,061 combined installs across 200 listings; 240 stars; MIT; the largest single-owner Claude skill collection in this batch, 187 SKILL.md files across 145 plugin directories
  • Adobe Skills - PDF, Document and Creative Suite Setup - 76,451 combined installs across 137 listings; 318 stars; Apache-2.0; Adobe product skills for AEM, Creative Cloud, Commerce, Analytics and the document skills, 178 SKILL.md files in per-product plugins
  • Jwynia Agent Skills - Presentation and Writing Suite - 65,356 combined installs across 119 listings; 180 stars; no LICENSE file at root; presentation design, writing and research skills from a single author, one directory for every indexed listing
  • ByteDance MediaKit CLI - Video Processing Setup - 61,321 combined installs across 5 listings; 218 stars; MIT; ByteDance MediaKit as five skills over a CLI for audio, video, image and editing pipelines, all five inside 38 installs of each other
  • DeerFlow Skills - Deep Research Agent Setup - 60,292 combined installs across 26 listings; 84,743 stars; MIT; the skills that ship inside the ByteDance deep research harness, most of them for working with the harness itself
  • LinkFox Skills - IP, Legal and Trade Data Setup - 58,424 combined installs across 200 listings; 115 stars; MIT; cross-border ecommerce and IP data skills for patents, trademarks, Amazon, Ozon and 1688, 270 SKILL.md files with Chinese descriptions throughout
  • Application Skills - SaaS Integration Suite Setup - 57,985 combined installs across 200 listings; 270 stars; no LICENSE file at root; Membrane managed-integration skills, 3,074 flat directories of about 173 lines each, every one requiring a Membrane account
  • Patricio Skills - Full Stack Engineering Suite Setup - 56,693 combined installs across 153 listings; 63 stars; MIT; 155 markdown skills in ten discipline folders from scaffolding through animation, API design, CI/CD, testing and AI engineering
  • Apollo GraphQL Skills - GraphQL and Rust Setup - 56,575 combined installs across 15 listings; 123 stars; MIT; official GraphQL skills for Apollo Client, Federation, Router, GraphOS and schema work, plus a general Rust style guide at the top of the install table
  • Patterns.dev Skills - UI and Rendering Patterns Setup - 53,715 combined installs across 58 listings; 258 stars; no LICENSE file at root; design and rendering pattern cards for JavaScript, React and Vue, one directory for every indexed listing, idle since April 2026
  • Scenario Labs Skills - Custom Model Training Setup - 53,338 combined installs across 173 listings; 1,284 stars; MIT; game and creative asset generation skills over the Scenario platform, 174 skill directories including a Blender expert and consistency tooling
  • Android Skills - Play Policy and App Quality Setup - 48,501 combined installs across 35 listings; 10,111 stars; Apache-2.0; Google's official Android skills for Play policy, edge-to-edge, navigation, R8 and app quality, 36 SKILL.md files across 18 domain folders
  • Lev Nikolaevich Skills - DevOps and Delivery Setup - 45,336 combined installs across 198 listings; 585 stars; MIT; DevOps and delivery skills numbered ln-1xx to ln-8xx inside plugin suites, 41 SKILL.md files against 198 indexed listings

New - October 30, 2026 Morning Sweep (19 publisher guides)

Nineteen repository-scoped clusters drained from the queue the October 29 evening probe left behind, for 345,882 combined installs across 264 indexed listings:

  • Remotion Framework Skills - Programmatic Video Setup - 47,489 combined installs across 73 listings; 69,018 stars; NOASSERTION while LICENSE.md carries the two-tier Remotion License; the React video framework's own skill library: 13 public skills around a router plus 63 internal maintainer skills, 88 SKILL.md files in three layouts, and four indexed names with no directory at main
  • Auth0 Agent Skills - Identity and OAuth Setup - 28,077 combined installs across 47 listings; 55 stars; Apache-2.0; official Auth0 identity skill consolidated into a single skill and a 144-file reference set, against 47 indexed per-framework names that are the retired ones
  • Upstash Context7 - Versioned Library Docs Setup - 25,472 combined installs across 8 listings; 63,945 stars; MIT; Context7 documentation retrieval as four skills over the cx7 CLI and the remote MCP endpoint, master branch, with three listings that have no directory in the tree
  • Svelte AI Tools - Official Svelte and SvelteKit Setup - 24,570 combined installs across 2 listings; 335 stars; MIT; the official Svelte MCP in two skills, svelte-code-writer and svelte-core-bestpractices, shipped across five harness layouts with a static autofixer CLI
  • Prisma Next Skills - Typed Database Access Setup - 23,921 combined installs across 40 listings; 411 stars; Apache-2.0; Prisma Next as 13 published skills plus 18 contributor skills, with migration review, upgrade walkthroughs and per-transition codemods
  • Axiom Skills - Observability and Event Queries Setup - 23,872 combined installs across 15 listings; 17 stars; MIT; official Axiom observability skills for APL queries, dashboards, alerting, SPL conversion and cost control, where v1.2.0 renamed the skills with an axiom- prefix
  • Lavish AXI - Agent Experience Interface Setup - 23,474 combined installs across 3 listings; 4,358 stars; MIT; one agent-facing UI skill backed by a CLI: a design vocabulary, hook and plugin targets, and a publish path to a hosted page
  • Multi-Agent Review - Adversarial Code Review Setup - 19,509 combined installs across 1 listing; 82 stars; no licence file; a four-adversary review protocol with a synthesis pass that turns competing critiques of a spec or plan into one decision list
  • Pulumi Agent Skills - Infrastructure as Code Setup - 16,370 combined installs across 17 listings; 78 stars; Apache-2.0; official Pulumi IaC skills in four plugin groups: authoring and ESC secrets, five migration converters for Terraform, CDK, CloudFormation, ARM and Bicep, provider maintenance and Neo handoff
  • components.build - UI Component Standards Setup - 14,813 combined installs across 1 listing; 802 stars; NOASSERTION while the LICENSE file is Apache-2.0 and the README says MIT; the components.build specification as a skill: 16 spec pages, 19 languages of translations and an accessibility reference set
  • Kapso Agent Skills - WhatsApp Business Setup - 13,788 combined installs across 9 listings; 195 stars; no licence file; Kapso WhatsApp Business skills for integration, automation, calls and observability, master branch, with five indexed listings that have no directory
  • Dotneet Claude Code Marketplace - Plugin Collection Setup - 12,917 combined installs across 10 listings; 3 stars; no licence file; a six-plugin marketplace for review, prompt improvement and agent session work, where call-claude fails Snyk and the guide leads with it
  • Swift Concurrency Skill - Structured Concurrency Setup - 12,209 combined installs across 1 listing; 578 stars; MIT; twostraws' Swift 6 concurrency expert: tasks, actors and isolation, with 12 reference files covering hotspots, diagnostics and migration
  • Vectorize Hindsight Skills - Agent Memory Setup - 11,470 combined installs across 15 listings; 56,373 stars; MIT; agent memory as 20 skills plus a first-class Hermes memory-provider plugin, with self-hosted and cloud deployment paths
  • Amazon Devices Agent Skills - Fire TV and Vega Setup - 10,978 combined installs across 2 listings; 6 stars; MIT; Amazon's own Fire TV and Vega skills: a five-phase multi-TV migration with templates and a validation checklist, plus React Native TV UI rules
  • Vinta Django AI Plugins - Django and Celery Setup - 8,390 combined installs across 5 listings; 166 stars; no licence file; Django and Celery experts with safe-migration and reviewer skills from a Django consultancy, five skills over ten SKILL.md files
  • Vercel Chat SDK Skills - Chat Bot Development Setup - 6,118 combined installs across 3 listings; 2,434 stars; MIT; the Chat SDK and its adapter-writing workflow, with four byte-identical SKILL.md copies and a published agent-skills index
  • Better Icons - Icon Search Skill Setup - 6,047 combined installs across 1 listing; 1,445 stars; MIT; icon search and selection as one skill over a CLI with preferences, history and a library-wide naming convention
  • Google ADK Docs - Agent Development Kit Setup - 16,398 combined installs across 11 listings; 1,530 stars; Apache-2.0; Google's Agent Development Kit documentation as six skills covering the dev guide, cheatsheet, evaluation, deployment, streaming and observability, where the indexed names outrun the tree

New - October 29, 2026 Evening Sweep (12 publisher guides)

Twelve publisher clusters found by a widened 47-query topic probe after the standing window (top row 295) and the trending board (600 entries, 77 sources, all covered) both came back empty, for 821,254 combined installs across 372 indexed listings:

  • Elastic Agent Skills - Elasticsearch and Kibana Suite - 128,923 combined installs across 54 listings; 603 stars; Apache-2.0; the official Elastic suite where 62 in-tree files are 31 skills mirrored across two trees, and 23 indexed names were removed in v0.6.0 while still carrying install counts
  • claude-mem - Persistent Memory for Coding Agents Setup - 101,470 installs across 34 listings; 104,989 stars; Apache-2.0; session history compressed into a searchable memory store, with search, explore, plan and report skills on top of it
  • Claude Ads - Paid Media Skill Suite Setup - 98,745 installs across 34 listings; 10,586 stars; MIT; a conductor plus per-platform specialists for Meta, Google, TikTok, X, Amazon and Apple, with the Snyk warns on the skills that touch live ad accounts
  • AI Research Skills - 100-Skill ML Research Library Setup - 97,018 installs across 100 listings; 14,111 stars; MIT; 98 atomic ML research skills across eleven numbered topic directories, installed a category at a time by a CLI whose positional argument matches category ids
  • K2F - Document Generation Skills Setup - 62,502 installs across 42 listings; 3 stars; Apache-2.0; semantic JSON compiled into Word, PowerPoint, fillable PDF and IDML, with a schema, a template catalog and a pack verifier behind the core skill
  • Roblox Brain - Roblox Studio Development Skills Setup - 61,628 installs across 37 listings; 143 stars; MIT; 29 Roblox Studio skills covering Luau, architecture, building, lighting, animation, GUI, economy, testing and publishing, with a FAIL on the testing listing the guide leads with
  • Bailian CLI - Alibaba Model Studio Agent Skills Setup - 59,841 installs across 9 listings; 366 stars; Apache-2.0; Alibaba Cloud's own bl CLI taught by nine skills for generation, fine-tuning, managed agents, web search and knowledge bases, documented from INSTALL.md because the README carries no skills commands
  • Coralogix cx CLI - Observability Agent Skills Setup - 58,206 installs across 32 listings; 233 stars; Apache-2.0; 19 end-user observability skills plus four contributor-only skills under .claude/skills/ that must never be presented as product skills; master branch
  • Sent Plugin - Messaging and Verification Skills Setup - 54,176 installs across 21 listings; 49 stars; MIT; SMS, WhatsApp and RCS messaging, compliance and migration in 84 files that are the same 21 skills across four adapter layouts, with a 27-tool remote MCP endpoint
  • Base44 Skills - AI App Builder Platform Suite Setup - 51,043 installs across 6 listings; 92 stars; MIT; SDK, CLI, remote dev, sandbox and production troubleshooting for the platform, on pre-release manifests with one plugin's skills dir shipped as a git symlink
  • SwiftUI Agent Skill - Apple UI Development Setup - 37,853 installs across 2 listings; 3,791 stars; MIT; the batch's only Production-tier guide, all verdicts Pass: one expert skill with 33 reference files and a checklist, plus an Apple docs rescan skill under a soft-deprecation rule
  • Slack CLI Skill - Agent-Driven Workspace Automation Setup - 9,849 installs across 1 listing; 0 stars; MIT; a Go CLI giving agents JSON reads and writes on a workspace without a resident MCP server, from the publisher behind the Groundcover and incident.io CLI guides

New - October 29, 2026 Morning Sweep (10 publisher guides)

Ten publisher clusters documented from the queue the previous fire left behind, plus one find the trending board understated 66x, for 109,244 combined installs across 14 indexed listings:

New - October 28, 2026 Evening Sweep (12 publisher guides)

New - October 28, 2026 Morning Sweep (9 publisher guides)

New - October 27, 2026 Evening Sweep (9 publisher guides)

New - October 27, 2026 Morning Sweep (5 publisher guides)

AI answers on your data. See where every number comes from.

Connect your tools and ask in ChatGPT, Claude, or Perplexity. Enterprise-grade answers in minutes. From $29.95/month.

Start free, no credit card
Contact Terms Privacy Pricing