Agent Ecosystem Atlas

A schema-first, contributor-ready map of AI agent frameworks, SDKs, skills systems, and agent tooling.

_{If this repository helps you, starring and sponsoring directly support ongoing curation.}

---

Why This Project

The AI agent ecosystem is moving fast. Decision quality falls when catalogs are inconsistent, promotional, or hard to maintain.

This repository is built to solve that with:

• a stable, explicit classification system
• structured YAML entries with validation
• reproducible generated indexes
• upstream skill repository indexing
• local operator skills for curation workflows
• contributor onboarding for both beginners and advanced maintainers

At A Glance

Signal	Value
Repository entries	20 validated entries
Canonical taxonomy	15 top-level categories
Upstream skill sources	3 tracked repositories
Local operator skills	6 curation skills
Generated indexes	5 browse surfaces
Validation status	Script-validated in CI-ready workflow

Why Star This Repo

• It gives a strict, comparable schema for evaluating fast-moving agent tooling.
• It tracks upstream skill ecosystems without becoming a copy or fork.
• It stays contributor-friendly with templates, automation, and principles-first review rules.

Quick Navigation

• Quick Start
• Who This Is For
• Repository Structure
• Classification System
• Skills System
• Upstream Skill Repositories
• Adoption Prompt
• Data Schema
• Principles
• Indexes
• Contributing
• Roadmap

Quick Start

1. Install dependency

pip install -r requirements.txt

2. Validate data

python scripts/validate_repo_entries.py

3. Regenerate indexes

python scripts/generate_indexes.py
python scripts/index_upstream_skills.py

Who This Is For

Role	What You Get
Tool evaluators	Category-consistent entries with strengths and limitations
Engineers	Fast browsing by category, language, and use case
Beginners	Guided wiki docs and learning paths
Maintainers	Validation scripts, templates, and CI-ready structure
Contributors	Clear contribution flow with review checklist

Repository Structure

.
|-- README.md
|-- LICENSE
|-- CONTRIBUTING.md
|-- CODE_OF_CONDUCT.md
|-- SECURITY.md
|-- SUPPORT.md
|-- .github/
|-- docs/
|-- wiki/
|-- data/repositories/
|-- skills/
|-- templates/
|-- scripts/
|-- indexes/
`-- examples/

Key Paths

Path	Purpose
data/repositories/	Canonical YAML catalog entries
data/upstream/	External skill repository source manifests
skills/	Local operator skills for repo ingestion, classification, and comparison
indexes/	Generated browse views for discoverability
docs/	Core standards (taxonomy, schema, system design)
wiki/	Beginner-friendly conceptual learning content
templates/	Contributor templates for entries, upstream sources, skills, issues, and PRs
scripts/	Validation + index generation automation
.github/	Native GitHub issue/PR UX and CI workflow

Classification System

Every entry uses exactly one canonical category from a fixed 15-category taxonomy.

This prevents drift and enables clean comparison over time.

Read: docs/classification-system.md

Skills System

This repository treats external skill repos as primary sources.

Local skills/ are operator tools for:

• extracting skill patterns from upstream repos
• classifying and summarizing repositories
• comparing frameworks and skill systems

Local operator skills live in:

skills/<skill-name>/SKILL.md

Required frontmatter fields:

• name
• description
• category
• inputs
• outputs
• tags
• difficulty

Read: docs/skills-system.md

Upstream Skill Repositories

The project tracks upstream skill ecosystems via:

• data/upstream/skill-sources.yaml
• templates/upstream-skill-source-template.yaml

Browse source index:

• Upstream Skill Sources

Regenerate upstream index:

python scripts/index_upstream_skills.py

Adoption Prompt

Need another AI to apply this system to an existing repository?

• Use: docs/adoption-prompt.md

Data Schema

Repository entries include decision-critical fields:

• identity: name, github_url, description
• classification: category, subcategory, tags
• quality signals: maturity, activity_status, docs_quality, beginner_friendly
• comparison support: use_cases, notable_features, limitations
• maintenance: last_reviewed

Read: docs/schema.md

Principles

This project is run with explicit quality and stewardship principles.

• Read: PRINCIPLES.md
• Release history: CHANGELOG.md

Indexes

• By Category
• By Use Case
• By Language
• Beginner-Friendly

Beginner Path

Start here:

1. Beginner's Guide
2. Agent Types Explained
3. Choosing an Agent Framework
4. Learning Paths

Contributing

• Contributor guide: CONTRIBUTING.md
• Repo template: templates/repo-entry-template.md
• Upstream source template: templates/upstream-skill-source-template.yaml
• Skill template: templates/skill-template.md
• Issue template: templates/issue_template.md
• PR template: templates/pull_request_template.md

Roadmap

• Structured repository schema and taxonomy
• Reusable skills system
• Beginner wiki docs
• Automated validation and index generation
• GitHub-native issue/PR UX and CI workflow
• Expand entry volume with quarterly review cadence
• Add benchmark-backed comparison snapshots

Quality Principles

• facts over hype
• explicit limitations in every entry
• deterministic generation and validation
• clean separation of production data vs examples
• original content (not direct mirrors/forks of seed repos)