SkillFed
A search platform built for AI agents that indexes over 14,000 Python packages, 56,000+ agent skills, and 191 research papers, so agents can find current tools by describing what they need in plain language.
🔗 Visit SkillFed
Description
An AI model's training data has a cutoff date, so it doesn't know about the Python package someone published last week or the research paper that came out yesterday. SkillFed gives agents a way to search past that cutoff, indexing packages, skills, and papers continuously and letting agents query it in plain language.
SkillFed indexes 14,416 PyPI packages across 123 categories, 56,283 'skills' from 3,292 publishers, and 191 research papers from 2016 to 2026, refreshed daily (270+ arXiv papers, 180+ agent skills, 5,700+ package releases per update cycle). It's accessible via MCP integration, a JSON API, or direct Python installation, and lets agents search by capability description without needing to expose their code, rather than scrolling an overwhelming package list. Results include package metadata like license, maintenance status, and known vulnerabilities. It appears to be free, with no pricing disclosed.
💬 Our review
The short version: SkillFed addresses a real, specific problem — AI agents working with stale knowledge of available packages and tools — and its scale (14k+ packages, 56k+ skills) makes it more comprehensive than manually searching PyPI for this particular use case.
PyPI's own search and Libraries.io both let you search packages directly, but neither is built for agent consumption via MCP, nor do they index the broader 'agent skills' and research-paper layer SkillFed adds on top. GitHub's semantic code search covers a different angle — finding code rather than discovering packages/skills by capability — so it's complementary rather than a direct substitute. Since SkillFed doesn't publish pricing, it's worth confirming there's no usage cap before building critical agent infrastructure on top of it; as a discovery layer to wire into an agent's MCP toolset, it's low-cost to test given free access currently.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Aucune tarification publiée au moment de l'évaluation
Pros
14 416 paquets PyPI indexés sur 123 catégories
56 283 skills d'agents indexées
Mise à jour quotidienne (papers, skills, releases)
Accessible via MCP, API JSON ou installation Python
Cons
Aucune tarification publiée, modèle économique flou
Limité à l'écosystème Python/PyPI pour les paquets
Service jeune, fiabilité long terme à confirmer