EasyGlobe EasyGlobe
Back to Skills Hub

EasyGlobe Skills

AI & Machine Learning Skills Hub

A shared ai & machine learning category page for Claude Skills, Codex Skills, Gemini Skills, Kimi Skills, GLM Skills, and team workflows.

Claude Skills Codex Skills Gemini Skills Kimi Skills GLM Skills ChatGPT Skills

How to choose

Review the agents and workflows covered by this category, then open a skill detail page to check task boundaries, README notes, setup requirements, and expected outputs.

How To Use This Skill

Use this category to organize agent-ready skills for model evaluation, prompt systems, agent design, ML workflows, and AI product operations. Each page follows the same skills template so teams can compare categories quickly and adapt the workflow to Claude, Codex, Gemini, Kimi, GLM, or internal SOPs.

Usable AI Agents and Models

Claude SkillsCodex SkillsGemini SkillsKimi SkillsGLM SkillsChatGPT Skills

Works well as

Claude SkillsCodex SkillsGemini SkillsKimi SkillsGLM SkillsTeam SOPs

Curated Skill Resources

Useful Skill, MCP, Agent, and finance API resources imported from the FynnZhang designer tool library and manual curation.

Wind Agent
Agent A-shares

Wind finance agent with Feishu, WeChat, and DingTalk binding support.

Source: design.fynnzhang.space
Marvis
Agent Tencent

Tencent agent product with interesting scenario coverage.

Source: design.fynnzhang.space
agentic-in/inferoa
Loop Engineering Skills Loop engineeringAgent harnessToken optimizationInference

Inference-native tokenmaxxing agent harness for loop engineering, with slash-command loops, verification evidence, memory, context control, and routing visibility.

Source: github.com
thu-nmrc/openloop
Loop Engineering Skills Loop engineeringAgent workflowVerificationGuardrails

Agent-agnostic loop engineering framework for play-test-fix-verify-improve cycles with logs, heartbeats, baselines, guardrails, and auditable stop conditions.

Source: github.com
Claude Video
GitHub Weekly Trending Skills Video understandingTranscriptionFrame extractionMultimodal

Video-understanding skill that accepts URLs or local files, extracts scene-aware frames, collects captions or Whisper transcripts, and combines the evidence for agent analysis.

Source: github.com
Codex PPT Skill
GitHub Weekly Trending Skills PPTXPresentationsImage generationCodex

Generate visually unified, image-based PPT and PPTX decks from articles, reports, papers, notes, or outlines with Codex and other skill-compatible agents.

Source: github.com
Lineage Skill
GitHub Weekly Trending Skills Skill creationKnowledge distillationSource groundingCourses

Distill courses, books, videos, audio, PDFs, slides, transcripts, OCR, and notes into source-grounded mentor, expert, consultant, practitioner, or custom agent skills.

Source: github.com
Sprite Gen
GitHub Weekly Trending Skills Game spritesAnimation atlasImage generationQA pipeline

Generate clean 2D game sprites and animation atlases through a component-row pipeline with chroma-key cleanup, frame extraction, deterministic atlas composition, QA reports, and curation tools.

Source: github.com

Core Workflows

Context and input setup

Define the source material, constraints, goal, expected output, and quality bar for ai & machine learning work.

Agent execution workflow

Run the task through Claude Skills, Codex Skills, Gemini Skills, Kimi Skills, GLM Skills, or another agent with clear checkpoints.

Review and reusable handoff

Package the result into a checklist, reusable prompt, operating note, or implementation handoff.

Typical Outputs

Model evaluation plan
Prompt system spec
Agent workflow
Experiment checklist
Risk notes

Frequently Asked Questions

Can AI & Machine Learning skills be used with Claude, Codex, Gemini, Kimi, and GLM?

Yes. The category is written as a model-flexible skills hub entry, so the same workflow can become a Claude Skill, Codex Skill, Gemini Skill, Kimi Skill, GLM Skill, or team SOP.

Why use one template for every category page?

A shared template makes the directory easier to scan: every category shows supported agents, core workflow steps, expected outputs, and FAQs in the same structure.