sam/sns-network-solutions
1
0
Fork 0
Code Issues 31 Pull Requests 5 Actions Packages Projects Releases Wiki Activity
55a80ce6e3
sns-network-solutions/ai-core/knowledge/ai-skills-repos.md
Samuel James 55a80ce6e3 Move ai-core team into repo; relocate sns.md to ai-core/knowledge 
- Add ai-core/ — Chief, Scout, Scribe OpenClaw agents + shared knowledge base
- Consolidate research/reference docs and the sns.md brand foundation under
  ai-core/knowledge/
- Repoint every sns.md reference (business branding, divisions, root README) to
  ai-core/knowledge/sns.md
- Update root + ai-core READMEs to reflect the new structure
2026-07-17 15:33:58 -05:00

75 KiB
Raw Blame History

AI Agent Skills & Tools Repository

Purpose: This document is a curated resource list for AI agents and the humans who build them. Each entry includes a description written for machine consumption — an AI reading this should be able to determine which repo solves its current problem without visiting the link.

How to use this file: Search by category header or scan the Use-Case Index tables. Each table row gives you: the repo name (linked), what problem it solves, and when to reach for it.

Last updated: 2026-07-17 | Total repos tracked: ~130+

🏷️ Use-Case Index

🧑‍💻 Coding & Engineering

Use these when an agent needs to write better code, review code, enforce standards, or audit quality.

Repo When to Use What It Does
mattpocock/skills Before writing any feature code; for TDD, architecture review, or converting ideas to specs 60k+ stars. Composable skills: /tdd enforces test-first, /grill-me challenges assumptions, /diagnose finds root causes, /to-prd converts ideas to product specs, /caveman cuts token use by 75%. Works on any agent (Claude, Codex, Cursor).
amElnagdy/guard-skills After agent generates code, BEFORE committing Second-pass quality gates that catch 14 documented AI failure modes: hallucinated npm packages, swallowed exceptions, tests that assert nothing, fake success declarations. Includes clean-code-guard, test-guard, docs-guard.
cloudflare/security-audit-skill When you need a security review of a codebase or PR Orchestrates multiple parallel agents through 6 phases: recon → hunting → validation → reporting → structured output → independent verification. Produces machine-readable findings with real exploit impact ratings.
dmmulroy/skills When working in TypeScript projects Personal skills by Dillon Mulroy (author of better-result, 1.6k stars). Covers TypeScript software design patterns, code review checklists, refactoring strategies, and implementation workflows.
fivetaku/fablize When agents abandon tasks partway or skip verification Based on 19 A/B runs + 26 real sessions (~1,500 tool calls). Forces completion-as-procedure: agent must gather evidence, verify each step, and prove the task is done — not just claim success.
ronikobrosly/RigorLoop When agent output needs to be reviewed like a real PR Treats every agent work product as a reviewable engineering artifact: decisions written down, tests traced to contracts, design intent visible, validation evidence captured, review concerns attached.
ksimback/tech-debt-skill When you need a tech debt audit of an existing codebase Produces TECH_DEBT_AUDIT.md with executive summary, findings table (every finding has a file:line citation), top-5 priorities, quick wins section, and a "looks bad but is fine" section to prevent false alarms.
golbin/agent-skills When building PRDs or reviewing implementation against spec Two skills: prd generates evidence-backed living PRDs, review-implementation checks actual code against spec documents and flags drift.

🧠 Context Window & Token Management

Use these when agents are burning tokens re-reading files, losing context, or hitting usage limits.

Repo When to Use What It Does
Dubibubii/usage-limit-reducer When hitting Claude Code usage limits too fast Measures token usage from JSONL logs and applies 11 rules. Key insight: 98.5% of tokens go to re-reading conversation history, only 1.5% to generating. Targets the 98.5%.
nekocode/filetree-skill At session start in any repo, before the agent reads files Maintains FILETREE.md — one line per file with a content hash for staleness detection. Agent can grasp entire repo layout in ~200 tokens instead of thousands. Hash tells agent if file changed since last read.
mattpocock/skills /caveman When agent responses are too verbose and waste output tokens Activates 75% token reduction mode. Agent gives terse, information-dense responses. No filler, no restating the question.
Kulaxyz/self-learning-skills When agent keeps re-discovering the same deploy/debug/verify paths After a session where the agent solved something hard (5+ steps), this skill harvests the golden path into a reusable SKILL.md or AGENTS.md rule. Next session, agent loads the rule instead of re-discovering.
AI-Builder-Club/skills When you want agents that compound knowledge across sessions "Loop engineers" — agents that trigger autonomously, execute tasks, verify results, and log what they learned. The log prevents re-discovery and context waste on problems already solved.

🔄 Self-Improvement & Memory

Use these when you want agents that get smarter over time without manual configuration.

Repo When to Use What It Does
Kulaxyz/self-learning-skills After any complex multi-step session Watches for hard-won workflows (deploys, debugging sequences, verification paths) and harvests them into reusable SKILL.md or AGENTS.md instructions. Works with Claude Code, Cursor, any AGENTS.md-aware agent.
AI-Builder-Club/skills When building always-on autonomous agents Skills for "loop engineers": agents that trigger on schedule/event, pick up work, ship it, verify it passed, and log learnings. Work compounds without human prompting every step.
adridder/moltron When OpenClaw agents create bad skills that break or waste tokens Teaches agents to build skills properly, test them via OpenTelemetry, version them in Git, and self-heal when an API changes. One-line install: curl -sSL .../install.sh | bash then @moltron init.
openclaw self-improving-agent For any OpenClaw agent that should learn from corrections Every message: silently checks if user corrected it, if a tool failed, or if something new was learned. Logs one-liners to .learnings/. Daily cron promotes repeated patterns to permanent rules. Cost: effectively $0/day.

✍️ Writing & Anti-Slop

Use these when agent output reads like generic AI — repetitive, wordy, or full of LLM tells.

Repo When to Use What It Does
shreyashankar/plain-writing-skill When agent prose is bloated or uses jargon unnecessarily Forces plain-language writing rules. Includes a revision view showing exactly what changed and why. By Shreya Shankar (UC Berkeley ML researcher). Featured on Digg for fighting AI slop.
yetone/kill-ai-slop When agent output has low information density or predictable structure Detects and eliminates AI-generated low-quality patterns. Enforces: every sentence must contain genuine insight, no filler phrases, no predictable structure, high information-per-token ratio.
yzhao062/agent-style When writing technical docs, papers, READMEs, or commit messages "The Elements of Agent Style" — 21 rules (12 from Strunk/Orwell/Pinker + 9 field-observed LLM patterns). Tested: 45–82% reduction in AI-tell violations across Claude Opus, GPT-5.4, Gemini 3 Flash.
Anbeeld/WRITING.md Drop-in writing rules for any agent's system prompt Three sizes: Full (WRITING.md), Compact (~1200 words for AGENTS.md), Mini (~240 words for embedding in prompts). Calibrates voice to genre, avoids 30+ documented LLM tells. MIT licensed.

🎮 Game Development

Use these when building games with AI agents — they provide engine-specific knowledge agents lack.

Repo When to Use What It Does
gamedev-skills/awesome-gamedev-agent-skills Starting any game project with an AI agent 66 original, version-pinned skills + a master router. Describe what you're building; the router auto-loads the right engine skills (Godot, Unity, Unreal, web). Works on Claude Code, Cursor, Codex, Copilot, Gemini CLI.
majidmanzarpour/threejs-game-skills Building browser games with Three.js A "game director" skill that automatically routes to specialist sub-skills: gameplay mechanics, AAA-style graphics, UI systems, AI-generated 3D/image/audio assets, debugging, and release verification.
LayrKits/Sprite-Pipeline Converting video/animation to game-ready sprites Python pipeline: takes animation frames → produces clean 256×256 horizontal sprite strips. Includes an HTML viewer for previewing results.

🔒 Security

Use these when you need to audit code for vulnerabilities or enforce security standards.

Repo When to Use What It Does
cloudflare/security-audit-skill Full codebase or PR security audit Six-phase pipeline with multiple parallel agents. Each finding is independently verified by a separate agent before inclusion. Machine-readable output format.
amElnagdy/guard-skills Pre-commit check on AI-generated code Catches: hallucinated packages (npm packages that don't exist), auth bypass patterns, swallowed exceptions, defensive code that hides bugs, tests that pass without asserting anything. Run on the diff, not the whole codebase.
cloudflare/skills Building on Cloudflare platform Official skills: Workers, Agents SDK, Durable Objects, sandbox-sdk, wrangler, web-perf, MCP server building. Commands: /cloudflare:build-agent, /cloudflare:build-mcp.

🏗️ Architecture & Visualization

Use these when you need to create diagrams, maps, or structured specs from text/code.

Repo When to Use What It Does
cclank/lanshu-animated-architecture-diagram Creating technical architecture diagrams for presentations or docs Produces premium black-canvas diagrams with hand-drawn typography. Outputs: editable Excalidraw files, static PNG, and genuinely animated GIFs. Runs as a Codex skill with local renderer.
vladignatyev/brain-map-skill Visualizing a knowledge base or Obsidian vault Takes a folder of Markdown notes → produces one self-contained interactive HTML file: force-directed graph colored by theme, a timeline showing knowledge growth, click-to-inspect panels.
gokulrajaram/ProductSpec Before any agent starts building a feature Open standard for AI-readable software intent. Structured sections: Problem → Hypothesis → Scope → User Experience → Acceptance Criteria → Success Metrics. Plus Decision Trace for post-launch evidence. Adopted by Fortune 500 engineering teams.

🤖 Multi-Agent & Orchestration

Use these when coordinating multiple agents or launching new agents programmatically.

Repo When to Use What It Does
cft0808/edict When you need governance over multi-agent workflows 12 AI agents organized as Tang Dynasty bureaucracy (Three Chancelleries & Six Ministries). Mandatory review gates, permission-based routing, full audit trails. 14k+ stars. Built on OpenClaw.
openclaw/agent-skills When building workflows for OpenClaw agents Official canonical source for shared workflows: review closeout, remote validation, and other common patterns. Write once, reuse across all OpenClaw instances.
anthropics/launch-your-agent When a founder wants to ship a Claude Managed Agent from scratch Skill that interviews you, scopes a v0, launches in your account, grades performance, iterates, and sets up scheduling. No code required. Apache-2.0.

📊 Business & Startup

Use these for startup validation, brand building, or business operations with AI.

Repo When to Use What It Does
Kappaemme-git/codex-startup-pressure-test-skill Before building anything — validate the idea first Give it a startup idea, get back: verdict (build/don't), scorecard, core assumption to test, fatal flaws, problem reality check, competition landscape, first customer ICP + prospect list with buying signals, and a 7-day outreach plan.
arnabbagxd/Brand-building-skills When defining brand architecture for a company Defines how brands, sub-brands, and product lines relate to each other. Builds a system that scales without creating consumer confusion.
nateherkai/AIS-OS Solopreneurs or small teams starting with AI Three-skill kit: /onboard (personalized interview), /audit (assess current state), /level-up (build leverage week over week). Designed to compound over time.

🛠️ Tooling & Meta

Use these for managing, inspecting, or installing agent skills across platforms.

Repo When to Use What It Does
iamzhihuix/skills-manage When managing skills across multiple agent platforms Tauri desktop app (TypeScript + Rust). Single UI to view, install, sync, and organize skills across 20+ platforms: Claude Code, Cursor, Codex, Gemini CLI, Copilot, and more.
xigua-wang/skill-doctor When skills conflict, overlap, or behave unexpectedly Local-first inspector that shows: installed skills, which definition wins on conflict, trigger overlap, risky patterns (shell access, network, secrets), duplicate names. Scans OpenClaw, Codex, Claude, Cursor, Copilot.
tiangolo/library-skills When you want agent skills that stay current with library updates By Sebastián Ramírez (FastAPI creator). Installs skills that ship inside library packages via symlinks. When the library updates, the skill updates too. uvx library-skills (Python) or npx library-skills (JS/TS).
nekocode/filetree-skill At session start to give agent cheap repo orientation Maintains FILETREE.md: one-line description per file + content hash for staleness. Agent reads this first, then only reads files whose hash indicates they changed. Massive token savings on large repos.

🌍 Agent Perception & Computer Use

Use these when agents need to interact with the physical world, control desktops, or use browsers.

Repo When to Use What It Does
machinepulse-ai/world2agent When agents need structured real-world sensor data Open protocol (W2A): install a sensor → agent gets structured real-time signals. Any sensor, same schema. Enables proactive agents that react to the physical world.
actuallyepic/background-computer-use When an agent needs to control macOS without taking the user's pointer Local API for background desktop automation — controls native apps, browser windows, multi-window workflows while user keeps working.
browser-use/bux When agents need isolated browser sessions for web automation Browser Use Box — containerized sandboxed browsers for AI agent automation. Part of the browser-use ecosystem.
jasonkneen/openclicky When you want an AI buddy that sees your screen, talks, and points at things Open-source screen-aware AI assistant. Bundled skills: creative studio, Google Workspace, research, screen control. 377 stars.
openclaw/clawsweeper When maintaining repos with large issue/PR backlogs Official OpenClaw bot that triages 7,000+ issues. Reviews weekly, proposes closures with evidence, only acts when confidence is high.
deepseek-ai/TileKernels When optimizing GPU kernels for LLM inference Optimized LLM kernels built with TileLang (Python DSL). Automatic optimization, easy migration. By DeepSeek.

🏠 Self-Hosted Apps & Infrastructure

Use these as deployment targets, backend services, or tools to integrate with agent workflows.

Repo When to Use What It Does
coder/code-server When you need VS Code accessible from any browser VS Code running on a remote server, accessed through the browser. 70k+ stars. The standard self-hosted IDE. Pairs with agents that need a persistent dev environment.
docmost/docmost When you need a self-hosted wiki/docs platform Open-source Confluence/Notion alternative. Real-time collaborative editing, spaces, permissions, nested pages. AGPL-3.0. Docker-ready.
paperless-ngx/paperless-ngx When you want to digitize and search physical documents Scan, OCR, index, and archive documents into a searchable online archive. Tagging, full-text search, automated workflows.
scanopy/scanopy When you need auto-updating network documentation Scans your network, identifies hosts/services, generates interactive visualizations showing how everything connects. Rust. 5.1k stars. Runs as a daemon for continuous updates.
relaticle/relaticle When you need a CRM that agents can directly interact with Open-source CRM with 30 MCP tools built in. REST API, self-hosted. Agents can create contacts, tasks, deals via MCP protocol. Laravel + Filament.
kalamdb/KalamDB When you need a lightweight DB with per-user isolation for AI workloads Real-time, storage-efficient SQL database designed for multi-tenant AI applications. Per-user data isolation without the overhead of separate databases.
tess1o/geopulse When you need private location tracking (Google Timeline alternative) Self-hosted, privacy-first location timeline with automatic trip detection, Immich photo integration, and rich analytics dashboards.
tphakala/birdnet-go When you want AI-powered wildlife sound identification Realtime bird/bat/wildlife classifier using local AI inference. Runs 24/7 on a Raspberry Pi. Multi-model, multi-microphone support.
C4illin/ConvertX When you need a self-hosted file conversion service 1000+ format support. TypeScript/Bun/Elysia. Web UI for drag-and-drop conversion. Docker-ready.
release-argus/Argus When you want notifications about new software releases Lightweight monitor that watches GitHub/URLs for new releases and notifies via Gotify, Slack, webhooks, or other channels.
Grashjs/cmms When you need computerized maintenance management Self-hosted CMMS: work orders, preventive maintenance schedules, asset tracking, team collaboration. Web + mobile. Free.
slskd/slskd When you want a modern Soulseek client with web UI Client-server architecture for the Soulseek file-sharing network. Web interface, Docker support, YAML config.

Detailed Listings

Below are full entries organized by star count and date added.

⭐ Top Tier (10k+ stars)

mattpocock/skills

  • URL: https://github.com/mattpocock/skills
  • Stars: 60.2k ⭐
  • Description: "Skills for Real Engineers" — composable, small agent skills for daily engineering work. Not vibe coding.
  • Key Skills: /grill-me, /grill-with-docs, /tdd, /diagnose, /improve-codebase-architecture, /zoom-out, /to-prd, /to-issues, /caveman (75% token reduction)
  • Philosophy: Fixes 4 failure modes — misalignment, verbosity, untested code, and architectural decay. Based on DDD, TDD, and software design principles.
  • Compatibility: Claude Code, Codex, Cursor, any agent
  • License: MIT

🔥 Popular (1k–10k stars)

ConardLi/garden-skills

  • URL: https://github.com/ConardLi/garden-skills
  • Stars: 2.4k ⭐
  • Description: Curated collection of production-ready Agent Skills for Claude Code, Cursor, Codex, and other AI coding agents.
  • Skills: web-video-presentation, web-design-engineer, gpt-image-2, kb-retriever
  • Install: npx skills add ConardLi/garden-skills
  • License: MIT

skills-manage (iamzhihuix)

  • URL: https://github.com/iamzhihuix/skills-manage
  • Stars: 1.6k ⭐
  • Description: Tauri desktop app for managing AI coding agent skills across 20+ platforms (Claude Code, Cursor, Gemini CLI, Codex, etc.) from one place.
  • Stack: TypeScript + Rust (Tauri)
  • License: Apache-2.0

Cloudflare Skills

  • URL: https://github.com/cloudflare/skills
  • Stars: 1.4k ⭐
  • Description: Official Cloudflare agent skills for building on Workers, Agents SDK, Durable Objects, and the wider Cloudflare Developer Platform.
  • Skills: cloudflare (comprehensive platform), agents-sdk, durable-objects, sandbox-sdk, wrangler, web-perf, MCP server building
  • Commands: /cloudflare:build-agent, /cloudflare:build-mcp
  • License: Apache-2.0

fancyai-official/skills

  • URL: https://github.com/fancyai-official/skills
  • Stars: 1k ⭐
  • Description: Premium AI skills for fashion, e-commerce, beauty, and editorial photography. Real-world creative production workflows from LVMH Innovation Award winners.
  • Categories: E-Commerce (beverage, beauty, kidswear, footwear, appliances), Fashion & Editorial (campaigns, editorial, design, street style)
  • License: Apache-2.0

🟢 Notable (200–999 stars)

charlie947/social-media-skills

  • URL: https://github.com/charlie947/social-media-skills
  • Stars: 783 ⭐
  • Description: Complete Claude skills behind Charlie Hills' content system (350k+ followers, 100m+ views/year). Newsletter → all channels pipeline.
  • Skills: 17 skills including voice-builder, post-writer, post-scorer, reels-scripting, youtube-thumbnail, gemini-infographic, analytics-dashboard, niche-research
  • License: MIT

bradautomates/claude-video

  • URL: https://github.com/bradautomates/claude-video
  • Stars: 703 ⭐
  • Description: Give Claude the ability to watch any video. /watch downloads, extracts frames, transcribes (via yt-dlp + Whisper), and hands it all to Claude.
  • How it works: Extracts frames + transcript → Claude reads frames as images + timestamps → answers grounded in actual video content
  • License: MIT

tiangolo/library-skills

  • URL: https://github.com/tiangolo/library-skills
  • Stars: 431 ⭐
  • Description: Install agent skills that ship with libraries (FastAPI, Streamlit, etc.) — always up to date via symlinks to installed packages.
  • Install: uvx library-skills (Python) or npx library-skills (JS/TS)
  • Author: Sebastián Ramírez (FastAPI creator)
  • License: MIT

yzhao062/agent-style

  • URL: https://github.com/yzhao062/agent-style
  • Stars: 396 ⭐
  • Description: "The Elements of Agent Style" — 21 writing rules (12 canonical from Strunk/Orwell/Pinker + 9 field-observed LLM patterns) for technical prose.
  • Results: 45–82% reduction in AI-tell violations across Claude Opus, GPT-5.4, Gemini 3 Flash
  • Scope: API docs, design docs, research papers, READMEs, commit messages
  • License: CC BY 4.0 (rules) + MIT (code)

LayrKits/Sprite-Pipeline

  • URL: https://github.com/LayrKits/Sprite-Pipeline
  • Stars: 355 ⭐
  • Description: Video-to-sprite-sheet pipeline for turning animation frames into clean 256×256 horizontal sprite strips. Includes AI assistant skill.
  • Stack: Python + HTML viewer
  • Use case: Game dev, 2D animation

ksimback/tech-debt-skill

  • URL: https://github.com/ksimback/tech-debt-skill
  • Stars: 354 ⭐
  • Description: Claude Code skill that produces a thorough, file-cited tech debt audit. Not a generic checklist — forces orientation before judgment, requires file:line citations on every finding.
  • Output: TECH_DEBT_AUDIT.md with executive summary, findings table, top-5 priorities, quick wins, and "looks bad but is fine" section
  • License: MIT

xigua-wang/skill-doctor

  • URL: https://github.com/xigua-wang/skill-doctor
  • Stars: 290 ⭐
  • Description: Local-first inspector for coding-agent skills. Shows what's installed, which definition wins, where risk accumulates, and how a workspace behaves across scopes.
  • Detects: Precedence conflicts, trigger overlap, risky patterns (shell, network, secrets), duplicate names
  • Scans: OpenClaw, Codex, Claude, Cursor, Copilot, GitHub, generic skill roots
  • Stack: React UI + TypeScript core
  • License: MIT

nateherkai/AIS-OS

  • URL: https://github.com/nateherkai/AIS-OS
  • Stars: 229 ⭐
  • Description: AI Operating System starter kit for Claude Code. Three-skill kit (/onboard, /audit, /level-up) + 3Ms framework for solopreneurs and small business operators.
  • Philosophy: Personalizes via interview, then builds leverage week over week
  • License: MIT

Anbeeld/WRITING.md

  • URL: https://github.com/Anbeeld/WRITING.md
  • Stars: 207 ⭐
  • Description: AI writing rules that make generated text sharper and less generic. Calibrates voice to genre, keeps format appropriate, avoids common LLM tells.
  • Variants: Full (WRITING.md), Compact (~1200 words for AGENTS.md), Mini (~240 words for embedding)
  • License: MIT

intertwine/dspy-agent-skills

  • URL: https://github.com/intertwine/dspy-agent-skills
  • Stars: 208 ⭐
  • Description: Production-grade DSPy 3.2.x skills for coding agents. Five skills covering fundamentals, evaluation, GEPA optimizer, BetterTogether, and RLM.
  • Validated: Against real DSPy 3.2.0 API with end-to-end examples
  • License: MIT

golbin/agent-skills

  • URL: https://github.com/golbin/agent-skills
  • Stars: 180 ⭐
  • Description: Reusable agent skills for Codex — prd (evidence-backed living PRDs) and review-implementation (review code against spec docs).
  • License: Not specified

🟡 Emerging (Under 200 stars)

Dubibubii/usage-limit-reducer

  • URL: https://github.com/Dubibubii/usage-limit-reducer
  • Stars: 124 ⭐
  • Description: Claude Code skill that stops you from burning through your usage limit. Measures token usage from JSONL logs and applies 11 rules for cutting consumption.
  • Key insight: 98.5% of tokens go to re-reading history, only 1.5% to generating responses

rampstackco/claude-skills

  • URL: https://github.com/rampstackco/claude-skills
  • Stars: 116 ⭐
  • Description: 65 stack-agnostic Claude Skills covering the full website lifecycle: brand, design, content, SEO, dev, ops, growth, and research. Includes Ahrefs MCP-powered SEO audit suite.
  • License: MIT

kwakseongjae/oh-my-design

  • URL: https://github.com/kwakseongjae/oh-my-design
  • Stars: 98 ⭐
  • Description: One-command bootstrap for skill-driven design. Generates DESIGN.md from 67 real company design systems (Google Stitch tokens + brand philosophy layer).
  • Skills: omd:apply, omd:init, omd:harness, omd:remember, omd:learn, omd:sync
  • License: MIT

🔍 Could Not Verify (Truncated URLs)

These repos had truncated URLs in the source video. Descriptions are inferred from names:

Repo Likely Purpose
paper2code (PrathamLearnsToCod) Convert research papers to code implementations
web-design (KAOPU-XiaoPu) Web design agent skill
claudedesign-to-swiftui (heyadam) Convert Claude design output to SwiftUI code
design-council (sjsyrek) Design review/council skill
code-overhaul-skill (ehmo) Codebase overhaul/refactoring skill
agent-session-resume (hacktivist123) Resume agent sessions across context windows
agent-sprite-forge (0x0funky) Sprite generation for game dev
linus-torvalds-skills (leopiney) Linux kernel/systems coding style skill
og-image-skill (stevysmith) Open Graph image generation skill
mizchi/skills Agent skills collection (Japanese dev)
Third Brain V5 Skills (Mark393295827) Personal knowledge management skills
eridani-speak (SijuEC) Speech/voice synthesis skill
founder-playbook (getagentseal) Startup founder workflow skills
shots (hypersocialinc) Screenshot/visual capture skill
GodModeSkill (99xAgency) Unrestricted/power-user agent skill
compose-performance-skills (skydoves) Jetpack Compose performance optimization

Categories Summary

Category Repos
Engineering/Dev Workflow mattpocock/skills, tech-debt-skill, golbin/agent-skills, library-skills
Writing/Style WRITING.md, agent-style
Design/UI oh-my-design, garden-skills (web-design-engineer), Sprite-Pipeline
Social Media/Marketing social-media-skills, rampstackco/claude-skills
Platform-Specific Cloudflare Skills, dspy-agent-skills
Tooling/Meta skills-manage, skill-doctor, usage-limit-reducer
Business/Ops AIS-OS, fancyai-official/skills
Media claude-video

Installation Quick Reference

# Most skills install via npx
npx skills add <owner>/<repo>

# Or Claude Code plugin marketplace
/plugin marketplace add <owner>/<repo>
/plugin install <skill-name>

# Or manual copy
cp -r skills/* ~/.claude/skills/

Source: YouTube video reviewing AI agent skills repos, May 2026

🆕 July 2026 Additions (OpenClaw/Agent Skills Video #2)

Curated from a second YouTube video showcasing OpenClaw and AI agent skills repos. Researched 2026-07-17.

Official / Platform Skills

openclaw/agent-skills

  • URL: https://github.com/openclaw/agent-skills
  • Description: Official shared skills for coding agents working on OpenClaw projects. Canonical source for common workflows like review closeout and remote validation.
  • Philosophy: Write a workflow once, reuse it everywhere, avoid hand-copying long SKILL.md files.
  • License: Not specified

anthropics/launch-your-agent

  • URL: https://github.com/anthropics/launch-your-agent
  • Description: Claude Code skills that take a founder from idea to a live Claude Managed Agent — interview, scope a v0, launch in their own account, grade it, iterate, and schedule it.
  • Use case: Technical founders building on Claude Managed Agents (CMA) — internal workers, product features, or customer-facing agents.
  • License: Apache-2.0

Security & Quality Gates

cloudflare/security-audit-skill

  • URL: https://github.com/cloudflare/security-audit-skill
  • Description: A coding-agent skill that turns your agent into a security auditor. Orchestrates multiple parallel agents through a six-phase pipeline — recon, hunting, validation, reporting, structured output, and independent verification.
  • Output: Machine-readable findings with real exploitable impact
  • License: Not specified

amElnagdy/guard-skills

  • URL: https://github.com/amElnagdy/guard-skills
  • Description: Focused guard skills for coding agents — second-pass quality gates that catch 14 systematic failure modes of AI-generated code, tests, and docs before they ship.
  • Skills: clean-code-guard (SOLID, DRY, KISS, YAGNI + AI failures), test-guard (9 rules for test quality), docs-guard (verifies docs against actual source)
  • Use: Run after agent writes code, before commit/merge
  • License: Not specified

Self-Improvement & Learning

AI-Builder-Club/skills

  • URL: https://github.com/AI-Builder-Club/skills
  • Description: Skills for building "loop engineers" — agents that get triggered on their own, pick up work, ship it, verify it, and log what they learned so work compounds without prompting every step.
  • Focus: Codebase harness + autonomous loop engineering
  • License: Not specified

Kulaxyz/self-learning-skills

  • URL: https://github.com/Kulaxyz/self-learning-skills
  • Description: A self-improving skill for AI coding agents. Recognizes hard-won golden paths in a session and harvests them into reusable skills/rules for next time.
  • Compatibility: Claude Code, Cursor, any agent that reads AGENTS.md
  • License: Not specified

Prompting & Agent Behavior

kingbootoshi/directional-prompting

  • URL: https://github.com/kingbootoshi/directional-prompting (inferred from profile — also known for vibe-coding-prompts)
  • Description: Directional prompting techniques for coding agents. Author known for Codex delegation via tmux sessions and Claude Code orchestration.
  • License: Not specified

shreyashankar/plain-writing-skill

  • URL: https://github.com/shreyashankar/plain-writing-skill
  • Description: A plain-language writing skill for AI agents with a revision view that shows what changed. Makes agents write clearly without AI slop.
  • Author: Shreya Shankar (UC Berkeley ML researcher)
  • License: Not specified

fivetaku/fablize

  • URL: https://github.com/fivetaku/fablize
  • Description: A Claude Code plugin that makes Opus behave like Fable — enforces completion, evidence gathering, and verification as procedure rather than luck. Based on 19 A/B runs + 26 real sessions (~1,500 tool calls).
  • Key insight: On closed, answer-bearing work (code, logic, builds), the completion and evidence enforcement is what matters, not model choice.
  • License: Not specified

Rimagination/good-question

Visualization & Architecture

cclank/lanshu-animated-architecture-diagram

  • URL: https://github.com/cclank/lanshu-animated-architecture-diagram
  • Description: Codex skill and local renderer for premium black-canvas technical diagrams with hand-drawn typography, editable Excalidraw output, static PNG previews, and animated GIFs.
  • Author: lanshu (enterprise AI video toolkit creator — 411 prompts, 15 model notes, 7 Claude Skills)
  • License: Not specified

vladignatyev/brain-map-skill

  • URL: https://github.com/vladignatyev/brain-map-skill
  • Description: Turn a folder of Markdown notes (Obsidian vault or gbrain export) into one self-contained interactive HTML knowledge map — force-directed graph colored by theme, a timeline to watch growth, and click-to-inspect panels.
  • License: Not specified

nekocode/filetree-skill

  • URL: https://github.com/nekocode/filetree-skill
  • Description: Claude Code plugin that maintains FILETREE.md — a one-line description per file with content hashes for staleness detection. Lets the LLM grasp repo layout in a few hundred tokens.
  • License: Not specified

Game Development

gamedev-skills/awesome-gamedev-agent-skills

  • URL: https://github.com/gamedev-skills/awesome-gamedev-agent-skills
  • Description: 66 original, version-pinned game-dev skills plus a master router in SKILL.md format. Covers Godot, Unity, Unreal, web, and beyond.
  • How it works: Describe what you're building; the router loads matching engine and task skills automatically.
  • Compatibility: Claude Code, Cursor, Codex, Copilot, Gemini CLI
  • License: Not specified

majidmanzarpour/threejs-game-skills

  • URL: https://github.com/majidmanzarpour/threejs-game-skills
  • Description: Agent skills for building playable, polished Three.js browser games. Includes a director that routes gameplay, graphics, UI, asset generation, audio, debugging, and release verification.
  • License: Not specified

Multi-Agent & Orchestration

vannyben7/edict-agent (likely cft0808/edict)

  • URL: https://github.com/cft0808/edict
  • Description: Multi-agent orchestration inspired by Tang Dynasty's Three Chancelleries & Six Ministries bureaucracy. 12 AI agents (11 business roles + 1 compatibility role) with mandatory review gates, permission-based routing, and full audit trails. Built on OpenClaw.
  • Stars: ~14k ⭐ (combined with Rust port)
  • License: Not specified

amElnagdy/delegate-skills

  • URL: https://github.com/amElnagdy/delegate-skills (inferred from author profile)
  • Description: Task delegation skills for coding agents — offloading work to sub-agents. Author also maintains guard-skills.
  • License: Not specified

Workspace & Productivity

DannyMac180/skills

  • URL: https://github.com/DannyMac180/skills
  • Description: AI agent skills by Dan McAteer including codex-dynamic-workflows — plan and run supervised agent dynamic workflows with goal mode, subagents, approval gates, integration, verification, and reusable artifacts.
  • License: Not specified

gauss314/skills

dmmulroy/skills

  • URL: https://github.com/dmmulroy/skills
  • Description: Personal coding-agent skills for TypeScript software design, review, refactoring, and implementation workflows. Author of better-result (1.6k stars) and the Cloudflare skill.
  • License: Not specified

dzhng/skills

  • URL: https://github.com/dzhng/skills
  • Description: Agent skills by David Zhang — AI-powered research assistant creator known for iterative deep research combining search, web scraping, and LLMs.
  • License: Not specified

jakubkrehel/skills

  • URL: https://github.com/jakubkrehel/skills (profile)
  • Description: Skills collection including make-interfaces-feel-better (small design engineering details that compound into great interfaces) and oklch-skill (OKLCH color space for web).
  • License: Not specified

Business & Startup

Kappaemme-git/codex-first-customer-finder-skill

  • URL: https://github.com/Kappaemme-git/codex-startup-pressure-test-skill (related — same author)
  • Description: Codex skill that finds your first 10 customers. Returns a full report: early-customer verdict, ideal ICP, evidence-backed prospect shortlist with buying signals, personalized openers, and a 7-day outreach plan. Author also makes codex-sell-my-saas-skill and codex-design-audit-skill.
  • License: MIT

arnabbagxd/Brand-building-skills

  • URL: https://github.com/arnabbagxd/Brand-building-skills
  • Description: Brand architecture strategy skills — defines how a company's brands, sub-brands, and product lines relate, building a system that scales without confusion.
  • Skills: brand-architecture and related brand strategy workflows
  • License: Not specified

Tracing & Observability

Forsy-AI/forsy-trace-skill

  • URL: https://github.com/Forsy-AI/forsy-trace-skill (listed on skywork.ai/skillhub)
  • Description: Captures AI agent workflows as structured traces: task context, steps, tool use, observations, feedback, failures, retries, artifacts, outcomes, and learning signals.
  • License: Not specified

Airtable / No-Code

alexmcdonnell-airtable/hyperagent-public-skills

Education

temorozov/claude-tutors

  • URL: https://github.com/temorozov/claude-tutors (inferred)
  • Description: Tutoring skills for Claude — personalized learning, spaced repetition, and comprehension verification. Part of the growing tutor-skill ecosystem.
  • License: Not specified

🔍 Could Not Verify (Truncated/Missing URLs)

These repos had truncated URLs in the source video and could not be confirmed via search:

Repo Likely Purpose
alvinunreal/lazyskills Lazy/minimal agent skills (author runs awesome-openclaw, awesome-autoresearch)
mikubaka88/CCFA-Skills Unknown — possibly financial/compliance skills (CCFA = Certified Corporate Financial Analyst?)
xiongqi123123/awesome-rebuttal Academic paper rebuttal skills (similar to Review2Rebuttal ecosystem)
Cuimao777/eterna-image2image-skill Image-to-image generation skill (author has cuimao-translator)
YurunChen/repo-docs-skills Generate documentation from repositories
wy51ai/edulab Education-focused agent skills
hajimi-kun/latex-to-word LaTeX to Word conversion workflow skill

Source: YouTube video reviewing OpenClaw/AI agent skills repos, July 2026

🆕 July 2026 Additions — Batch 3 (Mixed Projects & Skills Video)

Curated from a third YouTube video showcasing a mix of agent skills, tools, and projects. Researched 2026-07-17.

Notable Projects

MengTo/Skills

  • URL: https://github.com/MengTo/Skills
  • Description: Agent skills collection by Meng To (Design+Code founder). Likely design and UI/UX focused skills.
  • License: Not specified

nasa/spacewasm

  • URL: https://github.com/nasa/spacewasm
  • Description: NASA JPL's open-source Rust-based WebAssembly 1.0 interpreter engineered for resource-constrained spacecraft embedded systems. Flight-compliant WASM runtime.
  • Category: Aerospace/embedded systems (not an agent skill, but a remarkable OSS project)
  • License: NASA Open Source

gokulrajaram/ProductSpec

  • URL: https://github.com/gokulrajaram/ProductSpec
  • Description: Open standard for AI-readable software intent. Markdown format with sections: Problem, Hypothesis, Scope, User Experience, Acceptance Criteria, Success Metrics. Plus Decision Trace for post-launch evidence.
  • Philosophy: Structure the parts machines must execute or compare — captures committed intent before implementation.
  • Author: Gokul Rajaram (tech investor, former Square/Facebook exec)
  • License: Not specified

peetzweg/opendisplay

  • URL: https://github.com/peetzweg/opendisplay
  • Description: Free, open-source Sidecar/Duet alternative — use iPhone or iPad as a true second monitor for Mac over USB or WiFi. Low latency H.264, Retina HiDPI, touch input.
  • Category: Developer tool (not an agent skill)
  • License: Not specified

Doriandarko/texts-to-transformer

Agent Frameworks & Runtimes

yusukebe/ax

  • URL: https://github.com/yusukebe/ax (note: distinct from ax-llm/ax and google/ax)
  • Description: By Yusuke Wada (Hono.js creator). Likely an agent experience tool. Author also has hono-skill (Agent Skill for Hono apps) and a personal skills repo.
  • Related: yusukebe/hono-skill — Agent Skill for developing Hono applications with inline API reference and request testing.
  • License: Not specified

RoboDojo-Benchmark/RoboDojo

Agent Skills & Coding Tools

yetone/kill-ai-slop

  • URL: https://github.com/yetone/kill-ai-slop
  • Description: Anti-slop skill/tool by yetone. Eliminates AI-generated low-quality patterns, ensuring output has genuine insight and high information density.
  • Related ecosystem: deslop (OpenClaw skill), stop-slop, anti-slop-writing, AI-SLOP-Detector
  • License: Not specified

ronak-create/FableCut

DannyMac180/fable-advisor

  • URL: https://github.com/DannyMac180/fable-advisor (truncated)
  • Description: Fable-method advisory skill by Dan McAteer (also maintains DannyMac180/skills). Related to the Fable agent methodology for completion and verification.
  • License: Not specified

Sahir619/fable-method

  • URL: https://github.com/Sahir619/fable-method
  • Description: Implementation of the Fable method for AI agent task completion — emphasis on evidence and verification as procedure.
  • License: Not specified

ronikobrosly/RigorLoop

  • URL: https://github.com/ronikobrosly/RigorLoop
  • Description: Treats agent work as a reviewable software engineering artifact. Decisions are written down, tests trace to contracts, design intent is visible, validation evidence is captured, and review concerns stay attached to changes.
  • Author: Roni Kobrosly (Head of Data Science at DrFirst, creator of causal-curve)
  • License: Not specified

ohad6k/ditto

  • URL: https://github.com/ohad6k/ditto (author also has VibeRaven)
  • Description: Related to the VibeRaven ecosystem — mission control for AI-built apps. Author builds architecture context skills and production evidence tracking.
  • License: Not specified

sunflower-of-parch.../codex-hygiene

eisenjimmy/Quant

Forward-Future/gpt-5-6-relay

AHepi/DeepReason

Design & UI

deepika-builds/liquid-glass

  • URL: https://github.com/deepika-builds/liquid-glass (truncated)
  • Description: iOS 26 Liquid Glass effect implementation. Part of the broader liquid-glass ecosystem for web/native.
  • Related: haider-nawaz/liquid-glass-skill (Claude Code skill for SwiftUI Liquid Glass)
  • License: Not specified

vinhhien112/Three.js-Object-Sculptor-Codex-Plugin

Utilities & Infrastructure

tinbase/tinbase

gnomeria/usbtree

gostonx/uninstally

thatmagicalcat/txm

kunchenguid/dotfiles

Productivity & Workflows

colebemis/notion-workers

KORAYTEACHER/fintech-advisor

HaochengLu/contextvc

CoscoAI/luxy-aisre

  • URL: https://github.com/CoscoAI/luxy-aisre
  • Description: AI SRE (Site Reliability Engineering) skill — likely infrastructure monitoring, incident response, or ops automation.
  • License: Not specified

Communication & Browser

Ravvioli/browsertabs-talking

Nanako0129/pilotfish

🔍 Could Not Verify (Truncated/Minimal Info)

Repo Likely Purpose
Robbyant/lingbot-world-v2 Chatbot/language bot world v2
withmarbleapp/os-taxonomy Operating system or open-source taxonomy/classification
xiaotianfotos/homerail Home automation or rail transport project
CWS6206/EasyLastSkill Simplified "last skill" pattern for agents
dk8827/ra-port Port/migration tool

Source: YouTube video reviewing mixed AI projects and agent skills, July 2026

🆕 July 2026 Additions — Batch 4 (Self-Hosted Apps & Tools Video)

Curated from a YouTube video showcasing self-hosted applications and developer tools. Researched 2026-07-17.

📄 Documentation & Knowledge

Repo URL Description
Docmost https://github.com/docmost/docmost Open-source collaborative wiki and docs — Confluence/Notion alternative. Real-time editing, spaces, permissions. AGPL-3.0.
paperless-ngx https://github.com/paperless-ngx/paperless-ngx Supercharged document management — scan, OCR, index, and archive physical documents into searchable online archive.
Lumen-Labs/brainapi2 https://github.com/Lumen-Labs/brainapi2 (truncated) Brain/knowledge API — likely a personal knowledge base or second-brain backend.

💻 Development & Code

Repo URL Description
code-server https://github.com/coder/code-server VS Code running in the browser on a remote server. 70k+ stars. The standard for self-hosted IDE.
rifkyputra/postlab https://github.com/rifkyputra/postlab API testing/development lab — likely a Postman alternative.
Studio-Saelix/sencho https://github.com/Studio-Saelix/sencho Unknown — likely a dev tool by Studio Saelix.

🗄️ Databases & Storage

Repo URL Description
KalamDB https://github.com/kalamdb/KalamDB Lightweight, real-time, storage-efficient SQL database. Per-user data isolation, designed for AI-era multi-tenant workloads.
tinbase/tinbase https://github.com/tinbase/tinbase Lightweight database or data storage tool.

🌐 Networking & Infrastructure

Repo URL Description
scanopy https://github.com/scanopy/scanopy Network diagrams that update themselves. Scans network, identifies hosts/services, generates interactive visualizations. Rust. 5.1k stars.
release-argus/Argus https://github.com/release-argus/Argus Lightweight release monitor — notifies of new software releases via Gotify/Slack/webhooks.
MertJSX/folderhost https://github.com/MertJSX/folderhost Likely a simple file/folder hosting server.
oomol-lab/open-connector https://github.com/oomol-lab/open-connector (truncated) Open connector — likely an integration/API connector framework by OOMOL.

📍 Location & Maps

Repo URL Description
geopulse https://github.com/tess1o/geopulse Self-hosted Google Timeline alternative — privacy-first location tracking with automatic trip detection, Immich integration, and rich analytics.

🎵 Media & Audio

Repo URL Description
slskd https://github.com/slskd/slskd Modern client-server app for the Soulseek file-sharing network. Web UI, Docker-ready.
birdnet-go https://github.com/tphakala/birdnet-go Self-hosted realtime bird/bat/wildlife sound identifier. Multi-model local AI, runs 24/7 on Raspberry Pi.
NeptuneHub/AudioMuse-AI https://github.com/NeptuneHub/AudioMuse-AI (truncated) AI-powered audio/music tool.
1146345502/aural-oss https://github.com/1146345502/aural-oss Open-source audio player/tool (Aural).

📊 Project & Work Management

Repo URL Description
Grashjs/cmms https://github.com/Grashjs/cmms #1 self-hosted CMMS (Computerized Maintenance Management System) — work orders, preventive maintenance, asset tracking, team collaboration. Free. 656 stars.
Gimanh/taskview https://github.com/Gimanh/taskview (truncated) Task view/management application.
wm94i/Work-Review https://github.com/wm94i/Work-Review Work review/performance management tool.
DaKheera47/job-ops https://github.com/DaKheera47/job-ops Job operations/application tracking tool.
famstack-dev/famstack https://github.com/famstack-dev/famstack Family/team stack management tool.

🤖 CRM & Business

Repo URL Description
relaticle https://github.com/relaticle/relaticle Open-source CRM with native AI agent support — 30 MCP tools, REST API, self-hosted. Built with Laravel & Filament.
ChatbotXIO/ChatbotX https://github.com/ChatbotXIO/ChatbotX Chatbot platform — likely a self-hosted conversational AI builder.
KORAYTEACHER/fintech-advisor https://github.com/KORAYTEACHER/fintech-advisor (truncated) Fintech advisory tool.

🔄 File Conversion & Utilities

Repo URL Description
ConvertX https://github.com/C4illin/ConvertX Self-hosted online file converter — supports 1000+ formats. TypeScript/Bun/Elysia.
gostonx/uninstally https://github.com/gostonx/uninstally Application uninstaller utility.
gnomeria/usbtree https://github.com/gnomeria/usbtree USB device tree viewer/manager.

💰 Finance & Budgeting

Repo URL Description
simplesuite/simplebudget https://github.com/simplesuite/simplebudget (truncated) Simple self-hosted budget tracker.
eisenjimmy/Quant https://github.com/eisenjimmy/Quant Quantitative finance/trading toolkit.

🏥 Health & Wellness

Repo URL Description
CodyBontecou/health-md https://github.com/CodyBontecou/health-md (truncated) Health tracking in Markdown — likely a self-hosted health data system.

🏠 Self-Hosted Misc

Repo URL Description
rejourneyco/rejourney https://github.com/rejourneyco/rejourney Travel/journey planning or journaling platform.
iuliandita/digarr https://github.com/iuliandita/digarr Digital media management (likely *arr-stack adjacent — Sonarr/Radarr style).
zymazza/mazzap https://github.com/zymazza/mazzap WhatsApp-related tool or automation.
shukiv/jabali-panel https://github.com/shukiv/jabali-panel Admin/management panel (Jabali).
Jstudner/jcorp-nomad https://github.com/Jstudner/jcorp-nomad Nomad/remote work management tool.
mhumayunsaeed/Corvus https://github.com/mhumayunsaeed/Corvus Unknown — likely a monitoring or management tool (Corvus = crow/raven).

🤖 AI & Agents

Repo URL Description
Clawdlinux/agentic-operator-core http://github.com/Clawdlinux/agentic-operator-core (truncated) Agentic operator core — likely an OpenClaw/agent runtime for Linux systems.
DiegoGuidaF/PulseWeaver https://github.com/DiegoGuidaF/PulseWeaver (truncated) Pulse/event weaver — automation or monitoring tool.

📱 Communication

Repo URL Description
tonyantony300/alt-sendme https://github.com/tonyantony300/alt-sendme (truncated) Alternative file/message sending tool.
peetzweg/opendisplay https://github.com/peetzweg/opendisplay iPhone/iPad as second Mac monitor — free Sidecar/Duet alternative.

Source: YouTube video reviewing self-hosted apps and tools, July 2026

🆕 July 2026 Additions — Batch 5 (Final — Mixed Agent Tools, Protocols & Projects)

Curated from a fifth YouTube video. Researched 2026-07-17. This is the final batch.

🤖 Agent Protocols & Perception

Repo URL When to Use What It Does
world2agent https://github.com/machinepulse-ai/world2agent When agents need real-world sensor data in a standard format Open protocol (W2A) that standardizes how AI agents perceive the physical world. Install a sensor, agent gets structured real-time data. Swap sensors freely — all speak the same schema. World event → Sensor → Signal → Agent → Action.
background-computer-use https://github.com/actuallyepic/background-computer-use When an agent needs to control macOS apps without stealing the user's mouse Local macOS API for controlling native apps, browser windows, and multi-window desktop workflows in the background. Agent operates without taking over the pointer — user keeps working.
browser-use/bux https://github.com/browser-use/bux When agents need sandboxed browser environments Browser Use Box — containerized browser environments for AI agents. Part of the browser-use ecosystem (main repo has massive stars). Provides isolated, reproducible browser sessions for agent automation.
openclicky https://github.com/jasonkneen/openclicky When you want an AI assistant that can point at your screen and talk Open-source AI teacher/buddy that lives next to your cursor. Sees your screen, talks to you, points at things. Bundled skills: creative studio, Google Workspace, research reports, screen control. 377 stars. By Jason Kneen.

🧹 OpenClaw Ecosystem

Repo URL When to Use What It Does
clawsweeper https://github.com/openclaw/clawsweeper When maintaining a large GitHub repo with stale issues/PRs Official OpenClaw conservative maintenance bot. Scans all issues and PRs weekly, suggests closures with evidence, keeps backlog reviewed. Only closes when evidence is strong. Handles 7,000+ issues.

⚡ GPU & Performance

Repo URL When to Use What It Does
deepseek-ai/TileKernels https://github.com/deepseek-ai/TileKernels When building or optimizing LLM inference kernels Optimized GPU kernels for LLM operations built with TileLang — a Python DSL for high-performance GPU kernels. Easy migration, agile development, automatic optimization. By DeepSeek.
sweriko/ai4anim-webgpu https://github.com/sweriko/ai4anim-webgpu When running AI character animation in the browser WebGPU port of AI4Animation — neural network-based character animation running directly in the browser via WebGPU compute shaders.

🔧 Developer Tools & Utilities

Repo URL When to Use What It Does
hiteshchoudhary/gitbackup https://github.com/hiteshchoudhary/gitbackup When you need automated Git repository backups Git backup tool by Hitesh Choudhary (popular dev educator, 1M+ YouTube subscribers). Automates backing up repositories.
earthtojake/text-to-cad https://github.com/earthtojake/text-to-cad (truncated) When generating 3D CAD models from text descriptions Text-to-CAD generation — converts natural language descriptions into 3D CAD models.
nakagami/grdpwasm https://github.com/nakagami/grdpwasm When you need RDP (Remote Desktop Protocol) in the browser via WASM Go RDP client compiled to WebAssembly — access remote desktops directly from a browser without native clients.
alash3al/stash https://github.com/alash3al/stash When you need a simple key-value store or caching layer Lightweight key-value storage by alash3al (known for Go-based infrastructure tools).
sigcli/sigcli https://github.com/sigcli/sigcli When you need a CLI tool for Signal messenger Command-line interface for Signal messaging — send/receive Signal messages programmatically.
thatmagicalcat/txm https://github.com/thatmagicalcat/txm When you need a terminal multiplexer Terminal multiplexer — manage multiple terminal sessions.
kiwifs/kiwifs https://github.com/kiwifs/kiwifs When you need a lightweight filesystem Lightweight filesystem implementation (KiwiFS).

🎨 Creative & Animation

Repo URL When to Use What It Does
Jakubantalik/transitions.dev https://github.com/Jakubantalik/transitions.dev (truncated) When building smooth UI transitions/animations Web animation/transition library or showcase for developers.
EmGi96/TrailPrint3D https://github.com/EmGi96/TrailPrint3D When converting GPS trails or paths into 3D printable models Turns trail/route data into 3D-printable objects — physical souvenirs of hikes, runs, or journeys.
ronak-create/FableCut https://github.com/ronak-create/FableCut When editing video with AI assistance AI-powered video editing tool related to the Fable agent methodology.

🔐 Security & Privacy

Repo URL When to Use What It Does
thedrmaker/DrPasswords https://github.com/thedrmaker/DrPasswords (truncated) When you need a self-hosted password manager Password management tool — likely a self-hosted alternative to cloud password managers.
intave/intave https://github.com/intave/intave When you need anti-cheat or game server security Anti-cheat system — likely for Minecraft or game server integrity.

📡 Networking & Communication

Repo URL When to Use What It Does
chekusu/wanman https://github.com/chekusu/wanman When managing WAN connections or failover WAN management tool — likely multi-WAN load balancing or failover management.
russellromney/honker https://github.com/russellromney/honker When you need a lightweight notification/alert system Notification or alerting tool (honker = horn/alert).
saltpi/Aria.X https://github.com/saltpi/Aria.X When you need a download manager Aria-based download manager with extended features.

📊 Data & Analytics

Repo URL When to Use What It Does
machinarii/hypervibe https://github.com/machinarii/hypervibe When analyzing vibes/sentiment or building analytics dashboards Analytics or sentiment analysis platform.
systalyze/utilyze https://github.com/systalyze/utilyze When you need system utilization monitoring/analysis System utilization analysis tool — monitors and reports on resource usage.
mendrik-private/sqv https://github.com/mendrik-private/sqv When you need SQL visualization or query validation SQL query visualization or validation tool.

🌐 Web & Apps

Repo URL When to Use What It Does
stainlu/hermes-labyrinth https://github.com/stainlu/hermes-labyrinth When building maze/labyrinth games or procedural generation Hermes-themed labyrinth — likely a game or procedural maze generator tied to the Hermes Agent ecosystem.
wxtsky/byob https://github.com/wxtsky/byob When building your own customized tool/platform "Build Your Own Bot/Backend" — framework for creating custom bots or services.
evanklem/evanflow https://github.com/evanklem/evanflow When building workflow automation Personal workflow automation tool.
jameesy/foundry-vault https://github.com/jameesy/foundry-vault When managing secrets or credentials in a Foundry-based stack Vault/secrets management for Foundry ecosystem.
Beever-AI/beever-atlas https://github.com/Beever-AI/beever-atlas When you need AI-powered knowledge mapping AI atlas/knowledge mapping tool by Beever AI.
asekachov/Markie https://github.com/asekachov/Markie When you need a Markdown editor or renderer Markdown editing/rendering tool.
Amansingh-afk/milli.nvim https://github.com/Amansingh-afk/milli.nvim (truncated) When you need a Neovim plugin for millisecond-level operations Neovim plugin — likely for fast navigation or timing operations.

💡 Miscellaneous / Could Not Fully Verify

Repo URL Likely Purpose
tashfeenahmed/FreeLLMAPI https://github.com/tashfeenahmed/FreeLLMAPI (truncated) Free LLM API proxy — aggregates free-tier LLM endpoints into a single API
GammaLabTechnologies/harmonist https://github.com/GammaLabTechnologi.../harmonist (truncated) Audio/music harmonization tool or AI music assistant
thClaws/thClaws https://github.com/thClaws/thClaws OpenClaw-related project or tool
jay-sahnan/signal https://github.com/jay-sahnan/signal Signal processing or messaging tool
julianmemberstack/marko https://github.com/julianmemberstack/marko (truncated) Likely related to Marko.js framework or membership tooling

Source: YouTube video #5 — final batch of mixed agent tools, protocols, and projects. July 2026.

🆕 July 2026 Additions — Batch 6 (AI Models, SDKs & Agent Infrastructure)

Curated from a sixth YouTube video. Researched 2026-07-17.

🧠 AI Models & Research

Repo URL When to Use What It Does
MoonshotAI/Kimi-K2.5 https://github.com/MoonshotAI/Kimi-K2.5 When you need an open multimodal agentic model Moonshot's most powerful open-source model. Native multimodal (vision+text) with agentic capabilities. Built on ~15 trillion mixed tokens atop Kimi-K2-Base. Supports instant and thinking modes.
QwenLM/Qwen3-ASR https://github.com/QwenLM/Qwen3-ASR When you need multilingual speech recognition Open-source ASR by Alibaba's Qwen team. Supports 52 languages/dialects, language detection, timestamp prediction, music/song recognition. Models: 0.6B and 1.7B parameters.
deepseek-ai/DeepSeek-OCR-2 https://github.com/deepseek-ai/DeepSeek-OCR-2 When you need advanced document OCR with visual understanding "Visual Causal Flow" — novel encoder that dynamically reorders visual tokens based on image semantics. SAM + Qwen2 architecture optimized for document understanding, text extraction, and visual grounding.

🛠️ Agent SDKs & Platforms

Repo URL When to Use What It Does
github/copilot-sdk https://github.com/github/copilot-sdk When building apps powered by GitHub Copilot programmatically Official multi-platform SDK exposing the Copilot CLI engine. Define agent behavior — Copilot handles planning, tool invocation, file edits. TypeScript, Python, Go, .NET, Rust, Java. Production-tested runtime.
cloudflare/moltworker https://github.com/cloudflare/moltworker When you want OpenClaw running serverlessly on Cloudflare Workers Official Cloudflare deployment of OpenClaw. The Gateway runs inside a Cloudflare Sandbox SDK container instead of your own hardware. No Mac mini or VPS needed — runs at the edge.
ColeMurray/background-agents https://github.com/ColeMurray/background-agents When you need agents working asynchronously in the background Open-source background coding agent system inspired by Ramp's Inspect. Agents work on tasks while developers focus elsewhere. Also known as "Open-Inspect".
hanxiao/claudecode-telegram https://github.com/hanxiao/claudecode-telegram When you want to control Claude Code from Telegram Telegram bot bridge for Claude Code. Send messages from Telegram, get full Claude Code responses back. By Han Xiao (Jina AI founder).

📚 Agent Skills & Behavioral Guidelines

Repo URL When to Use What It Does
forrestchang/andrej-karpathy-skills https://github.com/forrestchang/andrej-karpathy-skills When agents make wrong assumptions, over-complicate, or don't push back Based on Karpathy's observations on LLM coding pitfalls. Four principles: don't assume, make surgical changes, surface inconsistencies, define verifiable success criteria. Drop-in CLAUDE.md / CURSOR.md.
sickn33/antigravity-awesome-skills https://github.com/sickn33/antigravity-awesome-skills When you need a curated directory of agent skills with install paths 27k+ stars. Searchable, installable skill library across planning, coding, debugging, testing, security, infra, product, and growth. Covers major AI coding assistants.
VoltAgent/awesome-openclaw-skills https://github.com/VoltAgent/awesome-openclaw-skills (truncated) When browsing the OpenClaw skills ecosystem 5,400+ skills filtered and categorized from the official OpenClaw Skills Registry. Organized by category for easy discovery.
unslop-xyz/noodles https://github.com/unslop-xyz/noodles When you need an anti-slop framework beyond single files By unslop.xyz — likely a comprehensive framework for eliminating AI-generated slop patterns across entire projects, not just individual outputs.

📄 Document Intelligence

Repo URL When to Use What It Does
kreuzberg-dev/kreuzberg https://github.com/kreuzberg-dev/kreuzberg When you need to extract text/data from any document format for RAG Polyglot document intelligence with a Rust core. Extracts text, metadata, images, tables from 91+ formats (PDF, Office, images). Available for Rust, Python, Ruby, Java, Go, PHP, TS, C#. CLI, REST API, or MCP server. Designed for RAG pipelines.

🎨 Visualization & Frontend

Repo URL When to Use What It Does
lukilabs/beautiful-mermaid https://github.com/lukilabs/beautiful-mermaid (truncated) When Mermaid diagrams need to look polished/beautiful Styling and enhancement layer for Mermaid.js diagrams — makes standard flowcharts, sequence diagrams, etc. visually appealing.
zarazhangrui/frontend-slides https://github.com/zarazhangrui/frontend-slides (truncated) When building presentation slides with web tech Frontend slide deck framework or generator.

🔧 Developer Tools

Repo URL When to Use What It Does
alecthomas/t https://github.com/alecthomas/t When you need a minimal task runner Minimal task runner by Alec Thomas (author of Kong CLI parser, Participle, Chroma syntax highlighter). Likely a lightweight Make/Just alternative.
lucasgelfond/zerobrew https://github.com/lucasgelfond/zerobrew When you want a zero-dependency package manager or installer Lightweight tool with zero dependencies — likely a minimal installer or bootstrap utility.
HugoRCD/evlog https://github.com/HugoRCD/evlog When you need structured event logging Event logging library by Hugo RCD (Nuxt ecosystem contributor). Likely structured, typed event logs for applications.
rudrankriyam/App-Store-Connect-CLI https://github.com/rudrankriyam/App-Store-Connect-CLI (truncated) When managing iOS/macOS apps from the command line CLI tool for Apple's App Store Connect API — manage apps, builds, TestFlight, metadata without the web UI.
dcramer/dex https://github.com/dcramer/dex When you need a developer experience tool By David Cramer (Sentry co-founder). Likely a DX-focused tool or framework.

🤖 Multi-Agent & Orchestration

Repo URL When to Use What It Does
yohey-w/multi-agent-shogun https://github.com/yohey-w/multi-agent-shogun (truncated) When orchestrating multiple agents in a hierarchy Multi-agent orchestration with a "shogun" (commander) pattern — hierarchical agent coordination.
Prismer-AI/Prismer https://github.com/Prismer-AI/Prismer When you need a multi-modal AI platform AI platform combining multiple modalities — likely vision, language, and action in one system.
mindfold-ai/Trellis https://github.com/mindfold-ai/Trellis When building structured agent workflows Trellis — likely a framework for structured, branching agent workflows (trellis = lattice structure).

📱 Apps & Productivity

Repo URL When to Use What It Does
Robbyant/lingbot-world-v2 https://github.com/Robbyant/lingbot-world-v2 (truncated) When building language learning bots Language learning chatbot platform v2. Same author has lingbot-depth for deeper language analysis.
Robbyant/lingbot-depth https://github.com/Robbyant/lingbot-depth (truncated) When you need deep language analysis Depth-focused language analysis companion to LingBot-World.
pierceboggan/primer https://github.com/pierceboggan/primer When you need a learning/onboarding tool Primer — likely a guided learning or onboarding platform for developers.
BradGroux/veritas-kanban https://github.com/BradGroux/veritas-kanban (truncated) When you need a kanban board with verification/truth features Kanban board with "veritas" (truth) focus — likely adds verification, evidence, or audit trails to task management.
chaitanyya/sales https://github.com/chaitanyya/sales When building sales qualification workflows Sales qualification tool (listed as "Qualify" in the video).

🔍 Could Not Fully Verify

Repo URL Likely Purpose
jmuncor/sherlock https://github.com/jmuncor/sherlock Investigation/debugging tool (Sherlock pattern)
SamNet-dev/conduit-manager https://github.com/SamNet-dev/conduit-manager (truncated) Pipeline/conduit management tool
samvallad33/vestige https://github.com/samvallad33/vestige Trace/history/vestige tracking tool
barrosyan/PINNeAPPle https://github.com/barrosyan/PINNeAPPle Physics-Informed Neural Networks application

Source: YouTube video #6 — AI models, SDKs, and agent infrastructure. July 2026. This is the final batch.