𓅂 AKHU Observer SDK — Enterprise AI evidence infrastructure adapter
Directory
Browse agent skills
Search by intent: design systems, testing, deployment, security, documentation, Azure, Supabase, React, or any workflow you want an agent to perform better.
Showing 39,601–39,630 of 129,642 skills
Page 1,321 of 4,322 · loading 30 at a time
Self-hostable multi-layer semantic caching for Node.js LLM applications
Instalador de la skill EMI para Claude Code. Instala solo la skill de referencia, no el repositorio EMI.
Font analysis utility for Sanity Studio — inspect OpenType metadata, variable font axes, and compare multiple fonts side by side
Typed client SDK for Featurebase — messenger, changelog widget, feedback widget, and surveys.
CLI for installing cross-agent subagent definitions from git repos, URLs, or local paths
Generic plugin SDK core runtime — framework-agnostic, no Tracker-specific types or Tracker HTTP client.
Skills for quant research with LLMs that don't fall for the traps nobody talks about. Lookahead bias, ATM detection, survivorship, transaction-cost realism — 10 skills + 11 reproducible evals showing 8 of them measurably improve Haiku output. Claude Code & Cursor compatible, MIT.
Android-only Expo Module (SDK 54) for fintech apps: listens to incoming SMS, classifies them (transaction / OTP / promo), parses banking & wallet alerts across 60+ South-Asian, Indian, Bangladeshi and GCC institutions, extracts OTPs, and exposes aggregati
Joomla 5+/6 extension development skill for Claude Code — components, modules, plugins, templates with modern MVC, PSR-4, DI, service providers.
Hallucination-Resistant Authorization Layer (HRA-L2) CLI Demo for x402-style agent payments
Peer coordination for independently opened Pi sessions with groups, artifacts, and mailboxes
Typed TypeScript client for Pulsar -- the headless broadcast engine bundled in Prism.
Vim Connect Application SDK — typed wrapper over coreSdk with collection-generated entity types
Five production-ready Claude Skill templates: one-pager writer, code reviewer, brand voice keeper, CSV cleaner, decision-doc generator. Each demonstrates a different SKILL.md pattern. Drop into Claude Code or Claude.ai. MIT licensed.
A reflex layer for every input on the web. Drop in SmartField / SmartForm / SmartText and get on-device autofill, contextual spellcheck, smart paste, and cross-field inference — zero latency, zero cost, zero privacy exposure.
A Claude Code skill that runs an end-to-end system design course — anchored to DDIA and the System Design Primer, with real-world incidents, runnable exercises, and mock interviews.
Git-backed package and profile manager for portable AI coding-agent capabilities.
Community AI SDK provider for Cursor SDK agents.
Core Extraction and Sanitization for Alete Edge
Automated QA testing agent with blueprint-based testing
Deprecated — use @sheepit-ai/sdk-js. This package re-exports @sheepit-ai/sdk-js for backward compatibility.
Editor-time lint rules for CEF agents. Each **agent** rule mirrors a check that `cef build` performs, so warnings in your editor agree with build failures in CI.
Claude-native, governed offensive-OSINT & external attack-surface platform for authorized red-team and bug-bounty. Multi-agent recon across web, identity, cloud, secrets & breach intel under deterministic scope/evidence/approval controls, with human-reviewed findings for hand-off to pen-testing. Not an exploitation framework.
Save the last assistant message to a markdown file via /save-output
PLATO — The Agent Training Platform. Other frameworks build agents. PLATO raises them.
Uranus Technologies AI components — chat, composer, streaming, audio, attachments, reasoning and planning surfaces built on @uranus-workspace/design-system and the Vercel AI SDK.