n8n community node plugin to optimize LLM token usage using Headroom context compression
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 24,631–24,660 of 128,199 skills
Page 822 of 4,274 · loading 30 at a time
TBook Vault server SDK — typed gateway REST client (accounts, tx intents, positions), webhook verification, and Sui decode-and-assert for Node.js backends
A ZCode/Claude skill that indexes Dart language features, core libraries, and tooling (analyzer/pub/devtools/FFI/JS interop) by Dart version — only Stable, never Experimental/Preview. Covers Dart 3.0 through 3.12. Install-and-use: copies itself into .clau
Any HuggingFace model. Local. Multi-modal. Served over an OpenAI-compatible API.
Parse .env files and detect/redact sensitive values. CLI + library, zero dependencies. Browser + Node.js compatible.
Global AI Skills Ecosystem for code review and audit
Visual Ultracode workflow dashboard for Claude Code CLI.
OpenGATE — Open Grounded AI Testing & Evaluation. An open-source framework for evaluating evidence-grounded AI systems.
A ZCode/Claude skill that indexes Kotlin language features, standard-library APIs, compiler capabilities, and platform (JVM/Native/JS/Wasm/KMP) tooling by Kotlin version — only Stable/GA, never Experimental. Covers Kotlin 1.3 through 2.4. Install-and-use:
A lightweight upstream Prompt Compiler for Vercel AI SDK / agent engineering. Compiles persona, system rules, and current input into system/prompt strings.
HTTP client with caching, request deduplication, typed mutations, SSE streaming, and interceptors
Kyo-so: MCP-native, ACP-powered multi-agent review gates for AI coding workflows.
Kyo-so: MCP-native, ACP-powered multi-agent review gates for AI coding workflows.
Get 50% OFF your first month of Cursor Pro using my verified referral link https://cursor.com/referral?code=TMB5HVRAHNX2 Active Cursor AI code editor discount code, promo code, and coupons for 2026.
Helix: a free, headless AI software-engineering agent and CLI. Runs a grounded loop (gather, act, verify, iterate) with permission-gated tools, snapshots, a per-task cost meter, and an eval harness. BYOK, provider-agnostic (Anthropic, Gemini, OpenAI-compa
命理圆桌会 · A multi-school Chinese BaZi (Four Pillars) roundtable reading skill for Claude
Save AI tokens: index your codebase for LLMs and cut token usage by up to 99%. One command creates a compact project map (repo map) that Claude, ChatGPT, Gemini, Cursor and Copilot read instead of exploring your code. Local, zero dependencies, auto-update
napi-rs native addon wrapping Meilisearch's milli search engine
napi-rs native addon wrapping Meilisearch's milli search engine
AI-powered terminal code reviewer for your staged git diffs, powered by Groq.
Comprehensive pentesting & security check skill for Pi coding agent - headers, ports, SQLi, XSS, OWASP Top 10
Fable 5.1 solution tracing. Prompting skills so Opus, GPT and open-weight models work like Fable 5.1.
AI SDK v6 provider for Pi Coding Agent (supports 15+ LLM providers)
Deterministic, config-driven multi-agent delivery pipeline (client -> orchestrator -> workers). Zero runtime dependencies.
Nemotron / vLLM provider helpers for the Vercel AI SDK: model builder with the thinking-budget request transform, capability inference, served-model catalog probing + selection, and omni multimodal attachment parts.
Framework-agnostic Vercel AI SDK ReAct streaming loop: text + reasoning + tools + metrics. The one agent loop shared by the CLI, web, and extension.
A Claude Code skill that turns Claude into Fable — an orchestrator that plans, decomposes, and delegates multi-step dev work to specialized subagents.
Harness-blind LLM proxy for Bun: routes every request to the cheapest capable model (Anthropic + OpenAI), caches responses, auto-updates pricing, tracks savings on a live dashboard, and runs a request/response plugin pipeline