THE SKILL LAYER FOR AI ASSISTANTS

Better work.
Built on skills.

GNCore is a curated library of reusable instructions that help AI assistants approach real tasks with structure, context, and clearer checks.

One library. Across workflows.
Powered by your assistant.
01 / CHOOSE
03 / CREATE
GN / CORE
research / synthesize↗
code / review↗
design / iterate↗
A system of instructions.
Not another autonomous agent.
403reusable skills
30domain packs
103specialist skills
12documented targets
01 / HOW IT WORKSGUIDANCE, NOT MAGIC

Give your assistant
a better playbook.

GNCore skills are instructions your AI assistant reads and follows. They help frame a task, choose an approach, and check the outcome. The assistant still does the work with the tools and permissions it has.

01↗

Find the right skill

Search by goal, keyword or domain. Every skill has a specific job.

search "clean data"
02↗

Install for your assistant

Choose a supported application target and workspace or user scope.

install sheet-clean
03↗

Ask for the outcome

Invoke a slash skill where supported, then give your context and criteria.

/sheet-clean
04↗

Review the evidence

Inspect what the assistant did. Skills support better process—not guaranteed results.

validate / review
03 / SKILL EXPLORERTHE ENTIRE CATALOG

A skill for the
next step.

Explore all 403 documented skills with real IDs, descriptions, install commands, and individual pages you can share.

04 / QUICK STARTFROM CLI TO CONTEXT

Start with
one command.

Search first, then install a skill into the assistant and scope you actually use. The example below uses the documented VS Code Codex target.

Download / installation guide ↗
GNCORE / CLI ↗

$ gncore search "clean data"

# choose a matching skill ID

$ gncore install sheet-clean \
--apps vscode-codex --scope workspace

# use the skill where slash invocation is supported

› /sheet-clean audit this workbook

05 / YOUR TOOLSSUPPORTED INSTALL TARGETS

Fits into the
way you work.

The catalog documents twelve application targets, including coding assistants and managed bundles. Target behavior varies; check each adapter before installation.

✳VS Code Chat / Copilot skills
✳VS Code Codex skills
✳VS Code Claude Code skills
✳Google Antigravity
✳GitHub Copilot Chat
✳Claude Desktop managed bundle
✳Claude Code
✳Cursor managed bundle
✳Windsurf managed bundle
✳Continue.dev managed bundle
✳Gemini CLI skills
✳OpenAI Codex CLI
Compare the 12 target keys ↗
BUILT FOR CLARITY

Helpful structure.
Human judgment.

Real instructions

Skills define approaches and review checks; they are not standalone tools or guaranteed solutions.

No website account

This catalog runs as a static Cloudflare Pages site. Browsing and searching require no sign-in.

Know the limits

Results depend on your assistant, its available tools, and your review. Sensitive work needs appropriate professional oversight.

THE CREATOR

Made by a builder.
For builders.

GNCore is created by Paladugu Ganesh Naidu, a developer exploring AI systems, developer tools, security and open-source engineering.

Meet the creator ↗
06 / FAQGOOD QUESTIONS

The basics,
answered.

GNCore is a library of reusable instructions. Here are the important boundaries before using it.

Is GNCore an autonomous AI agent?

No. GNCore provides skill instructions for an assistant to follow. It does not independently execute projects, connect to tools, or guarantee results.

Do I need to install all 403 skills?

No. Search the catalog, choose the relevant skill ID, and install only what you need.

What is the difference between workspace and user scope?

Workspace scope installs in the current project context where supported. User scope is intended for broader reuse across a user environment; the exact paths and behavior vary by assistant.

Does every assistant support slash commands?

No. Slash invocation is available only where the target assistant supports native skill invocation. Otherwise follow that target’s skill loading behavior.

What does Cloudflare host here?

Only the public, static GNCore product website and catalog data—not the Python CLI, an AI model, or an execution backend.

Can I export a skill for ChatGPT?

The supplied catalog documents one skill per ZIP using gncore export --app chatgpt. Actual import requirements depend on the ChatGPT experience and should be checked against package documentation.

WHAT COULD YOU DO WITH A BETTER PLAYBOOK?

Make the next
move count.

Find your skill ↗
✳