Skills

Browse AI skills and tools

The supporting stack is often what determines whether a model is usable in real work. This directory makes that layer easier to compare.

Normalized skill registry

Skills are treated as compatibility and operations infrastructure.

Official sources, install method, deployment model, permission posture, provider support, and works-with relationships are first-class fields so this can evolve into a real registry and routing layer later.

Results
13
Supported providers
7
Sources & trust
Official-first
Secrets · security

1Password CLI

Score
88

Practical secret access layer for agents and scripts that need secure credential injection without hardcoding.

Best for
Agent automation · Coding
Works with
agent runtimes · CI jobs · local developer machines
Install method
manual · hybrid
Supported providers
Any provider
Supported hosts
CLI · local scripts · CI environments
Last verified
Apr 8, 2026
Coding agent · coding

Codex CLI

Score
88

Terminal-native coding agent workflow for implementing features, refactors, and technical reviews from a real repo.

Best for
Coding · Agent automation
Works with
OpenAI models · Git repos · terminal-first dev workflows
Install method
npm · local
Supported providers
OpenAI
Supported hosts
CLI · Repo workspace
Last verified
Apr 8, 2026
Browser automation · browser

Playwright

Score
88

Reliable browser automation layer for agent actions, QA checks, scraping flows, and human-in-the-loop web tasks.

Best for
Agent automation · Research
Works with
any model · web-actuation stacks · QA automation
Install method
npm · local
Supported providers
Any provider
Supported hosts
Node apps · Python apps · browser testing environments
Last verified
Apr 8, 2026
Coding agent · coding

Claude Code

Score
87

Strong coding and refactor assistant with especially high-quality explanations and calm change planning.

Best for
Coding
Works with
Anthropic models · repo refactors · review-oriented workflows
Install method
npm · local
Supported providers
Anthropic
Supported hosts
CLI · Repo workspace
Last verified
Apr 8, 2026
Speech to text · extraction

OpenAI Whisper

Score
87

Battle-tested transcription layer for meetings, podcasts, multilingual notes, and voice-driven workflows.

Best for
Research · Agent automation
Works with
audio pipelines · meeting notes · local transcription stacks
Install method
pip · local
Supported providers
OpenAI · Local runtime
Supported hosts
CLI · Python apps · local pipelines
Last verified
Apr 8, 2026
Agent orchestration · workflow

LangGraph

Score
86

Graph-based runtime for durable agent workflows, branching logic, and stateful multi-step execution.

Best for
Agent automation · Research
Works with
OpenAI · Anthropic · Google
Install method
pip · hybrid
Supported providers
OpenAI · Anthropic · Google · Meta
Supported hosts
Python apps · Node apps · agent backends
Last verified
Apr 8, 2026
Automation · workflow

n8n

Score
86

Popular workflow builder for gluing APIs, LLM calls, and back-office tasks into practical automations.

Best for
Agent automation
Works with
SaaS connectors · LLM APIs · ops workflows
Install method
hosted · hybrid
Supported providers
OpenAI · Anthropic · Google · Cohere
Supported hosts
Hosted app · Self-hosted · browser UI
Last verified
Apr 8, 2026
Research search · search

SerpApi

Score
85

Search data API that helps agents gather current web signals with consistent, parseable results.

Best for
Research · Agent automation
Works with
research agents · market scans · evidence gathering
Install method
hosted · hosted
Supported providers
Any provider
Supported hosts
API backends · automation tools · research stacks
Last verified
Apr 8, 2026
Automation · workflow

Zapier AI Actions

Score
84

Fast path to business automation through prebuilt actions across SaaS tools and operational systems.

Best for
Agent automation
Works with
SaaS-heavy businesses · no-code ops teams · AI action layers
Install method
hosted · hosted
Supported providers
OpenAI · Anthropic · Google · Any provider
Supported hosts
Hosted app · SaaS integrations · browser UI
Last verified
Apr 8, 2026
Execution runtime · connector

E2B Sandbox

Score
82

Ephemeral execution environment for agent-generated code, notebooks, and dynamic analysis tasks.

Best for
Coding · Agent automation
Works with
generated-code flows · analysis notebooks · agent sandboxes
Install method
hosted · hosted
Supported providers
Any provider
Supported hosts
cloud runtime · agent backends · notebook flows
Last verified
Apr 8, 2026
Knowledge memory · memory

Obsidian Memory Vault

Score
82

Personal knowledge layer that turns agent memory into editable markdown instead of opaque hidden state.

Best for
Research · Agent automation
Works with
assistant continuity · markdown-based memory · human review loops
Install method
manual · local
Supported providers
Any provider
Supported hosts
desktop app · local markdown repos
Last verified
Apr 8, 2026
Vector retrieval · connector

Weaviate

Score
82

Vector database and retrieval layer for grounding answers, semantic search, and memory-augmented agents.

Best for
Research · Agent automation
Works with
RAG systems · memory-heavy agents · semantic search
Install method
hosted · hybrid
Supported providers
OpenAI · Anthropic · Google · Meta · Any provider
Supported hosts
Cloud · Self-hosted · Kubernetes
Last verified
Apr 8, 2026
Tool protocol · mcp

MCP Toolbox

Score
81

A practical toolbox around the Model Context Protocol for exposing actions, files, and resources to compatible agents.

Best for
Agent automation · Coding
Works with
OpenAI-compatible agents · Anthropic-compatible agents · tool servers
Install method
manual · hybrid
Supported providers
OpenAI · Anthropic · Any provider
Supported hosts
desktop agents · local runtimes · tool servers
Last verified
Apr 8, 2026