skillforge-agent

539 lazy-loading AI skills for Claude, Gemini, and Scientific agents

What is skillforge-agent?

skillforge-agent is a unified toolkit of 539 production-ready AI agent skills — merged from three specialized repositories into one clean package.

| Category | Count | Source | |----------|-------|--------| | Claude | 340 | Claude Code system prompts & agent prompts | | Scientific | 148 | Research, science, engineering & analysis skills | | Gemini | 51 | Gemini-exclusive agent skills | | Total | 539 | |

Zero token waste — skills load on demand

Skills are not preloaded. The index is tiny (~200 KB). Only the skill you ask for is read into memory — at the moment you call load(). No startup cost, no wasted tokens.

import time           # Your app starts instantly
load("eda")           # Only THIS skill is loaded, only when you need it

Install

Python

pip install skillforge-agent

Node.js

npm install skillforge-agent

Quick Start

Python

from skillforge_agent import load, search, list_skills, categories

# Load a skill — reads from disk only at this moment
skill = load("scientific-brainstorming")
print(skill.name)        # "scientific-brainstorming"
print(skill.category)    # "scientific"
print(skill.prompt)      # full skill content

# Use directly as a string
print(str(skill))

# Search skills
results = search("data analysis")
for r in results:
    print(r["name"], r["category"])

# List by category
gemini_skills = list_skills(category="gemini")

# See counts
print(categories())
# {"claude": 340, "scientific": 148, "gemini": 51, "total": 539}

JavaScript / TypeScript

const { load, search, listSkills, categories } = require('skillforge-agent')
// or: import { load, search, listSkills, categories } from 'skillforge-agent'

// Load a skill
const skill = load('exploratory-data-analysis')
console.log(skill.prompt)

// Search
const results = search('protein', 'scientific', 5)
results.forEach(r => console.log(r.name))

// Stats
console.log(categories())
// { claude: 340, scientific: 148, gemini: 51, total: 539 }

CLI

# After pip install
skillforge stats
skillforge search "data analysis"
skillforge search "agent" --category gemini
skillforge list --category scientific
skillforge load scientific-brainstorming

# After npm install -g
npx skillforge stats
npx skillforge search "protein"
npx skillforge load exploratory-data-analysis

Skill Categories

Gemini (51 skills)

General-purpose agent skills covering accessibility, data analysis, cloud tools, frameworks, APIs, and more.

skill = load("exploratory-data-analysis", category="gemini")
skill = load("accessibility-a11y")

Scientific (148 skills)

Research and science-focused skills: protein structure, molecular dynamics, bioinformatics, brainstorming, and more.

skill = load("scientific-brainstorming")
skill = load("molecular-dynamics")
skill = load("protein-structure-prediction")

Claude (340 skills)

Claude Code system prompts, agent prompts, tool descriptions, and skill templates.

skill = load("agent-prompt-explore", category="claude")

API Reference

load(name, category=None)

Load a skill by name. Returns a Skill object.

| Field | Type | Description | |-------|------|-------------| | .name | str | Skill name | | .category | str | gemini, scientific, or claude | | .description | str | One-line description | | .prompt | str | Full skill content |

search(query, category=None, limit=10)

Search skills by keyword. Returns a list of metadata dicts.

list_skills(category=None)

List all skills. Returns a list of metadata dicts.

categories()

Returns skill counts per category.

Project Structure

ai-skillforge/
├── skills/
│   ├── claude/          # 340 Claude Code system prompts
│   ├── gemini/          # 51 unique Gemini agent skills
│   └── scientific/      # 148 scientific research skills
├── index.json           # Unified skill registry (lazy-loaded)
├── python/              # Python package source
│   └── skillforge_agent/
│       ├── __init__.py
│       ├── loader.py
│       └── cli.py
├── js/                  # Node.js package source
│   ├── src/
│   │   ├── index.js
│   │   └── index.d.ts
│   └── package.json
└── README.md

Contributing

Pull requests welcome! To add a new skill, create a folder in the appropriate category:

skills/gemini/my-new-skill/SKILL.md

With frontmatter:

---
name: my-new-skill
description: One-line description of what this skill does.
license: MIT license
metadata:
    skill-author: YourGitHub
---

# Skill Title
...

Then regenerate index.json by running:

python3 scripts/build_index.py

License

MIT — Made with ❤️ by Lord1Egypt