Browse agent skills

A cache doctor for Claude Code: find the money your prompt cache is leaking.

@noveltyking/pdf-templates is a docs & writing skill indexed by Skillful.sh. Package: @noveltyking/pdf-templates.

A React-based visual CMS framework: headless page/component tree core with an optional Mantine-based visual editor

Claude Code skill: Fable-grade engineering discipline (explicit assumptions, scoped deliverables, evidence-backed reports) for any Claude model

Unified AI router runtime. Provider-agnostic entrypoint with OpenRouter protocol; OpenRouter plus local llama-cpp and transformersjs providers.

open-dmo — an aware AI DMO office in your terminal

Reusable framework for building voice AI web applications powered by Google Gemini Live and Firebase.

Run Pi coding tools in a Celesto computer

Benchmark AI coding agents (Claude Code, Codex, Gemini) against YOUR repo. Public benchmarks test them on someone else's code — AIBench Arena tests them on yours.

Playwright-friendly NodeMantis SDK

Shared pure core for the Fancy document/surface packages — the model an agent or human emits and edits. Flat node-tree + collab-safe fractional ordering, a typed op-spine contract (reduce + invert) with a reference tree reducer, an infinite-canvas primiti

Optional on-device whisper.cpp add-on for astramem-local's audio-file transcription capture (FEAT-489). Wraps the official ggml-org/whisper.cpp CPU-only CLI (bundled per-platform, v1: win32-x64 + linux-x64); GGML model downloaded + checksum-verified on fi

Text, Markdown, JSON-compatible, SARIF, annotation, and PR-comment renderers for Evidoc core reports.

Scheduled webhook notifier for the DDYS API with dedupe, retries, and multi-platform delivery.

Enterprise IT QA Documentation Generator and Project Initializer

Pi extension for Google NotebookLM — notebooks, sources, chat, research, and Studio artifacts via notebooklm-py CLI bridge.

n8n sub-node that chains multiple chat models (OpenAI, Anthropic, Mistral, Ollama/llama.cpp, ...) and automatically switches to the next model when one fails.

A Javascript-based SDK for integrating payment methods via 54Pay.

Pi extension that fans a prompt out to multiple LLM voices in parallel, then asks a distinct conductor model to synthesize the successful responses.

Exposure audit for vector search / RAG: coverage, dark matter, Gini, hub capture, CI gate

Cache-first DeepSeek coding agent for the terminal.

Domain-logic TypeScript library for IM-Juggling: intra-mart module catalog retrieval, repository access, project/module-structure operations, and WAR/static build output. ESM-only, Node >= 22.19.0.

Auto-commit extension for pi-coding-agent (checkpoint-then-reorganise strategy)

Semantic LLM-based per-turn model and reasoning router for Codex CLI

⚠️ POC — Dependency Confusion / Supply Chain Vulnerability. Originally an unclaimed npm package referenced in Vercel AI SDK documentation.

Layered context builder for LLM prompts: RAG + compressed session memory + user profile

Turn one topic into a finished Vox-style paper-collage explainer/ad video — automated end to end on Atlas Cloud + ffmpeg. An agent skill.

A from-scratch harness for agentic coding.

A CAP application-layer plugin for declarative, traceable, scheduled data pipelines between CAP services. Each pipeline is a linear READ -> MAP -> WRITE job between exactly one source and one target, with built-in tracker, retry, concurrency guard, manage

Rent real human browsers in n8n — anti-bot scraping, human-solved captchas, geo-targeting, and contract-based hiring of agents and humans. Powered by Ceki.