Agent Skills Hub

ai-llm-patterns

Anthropic Claude integration patterns: streaming, RAG with pgvector, tool use, model selection (Haiku/Sonnet/Opus), prompt caching, and cost management for AI-powered engineering.

reset

Resets workflow artifacts in the .otto/ directory. Safely removes tasks, specs, and browser sessions for a clean start.

btca-cli

Operate the btca CLI for source-first code research. Manage git, local, and npm resources to ground AI answers in actual codebase context rather than outdated documentation.

PAI

Physical hardware synthesis bridge for PAI. Generates blueprints, 3D printing code, SVG paths for laser cutting, and G-Code for CNC machining to bring agentic designs into the physical world.

hylo-ghl

GoHighLevel workflow automation expert. Integrates with Hylo GHL API to manage workflows, API endpoints, UI navigation, and automation planning.

skill-memory

A Git-backed memory store for agent skills. Download, version, edit, and share custom agent behaviors and procedural knowledge using a CLI.

ReasoningBank with AgentDB

Implement ReasoningBank adaptive learning with AgentDB's ultra-fast vector backend. Features trajectory tracking, verdict judgment, memory distillation, and pattern recognition for self-learning autonomous agents.

customer-research

Multi-source research tool for customer inquiries, bug investigations, and account history synthesis with source attribution and confidence scoring.

mijia-control

Control and monitor Xiaomi Mijia smart home devices including status switching, device discovery, automation scenes, and environmental statistics.

ios-debugger-agent

Automate iOS development workflows using XcodeBuildMCP: build, run, test, inspect UI, and capture logs on local simulators.

markitdown

Convert diverse file formats like PDFs, Office docs, images, audio, and web content into clean Markdown, specifically optimized for LLM ingestion, RAG pipelines, and automated text analysis workflows.

hs

Pre-execution security guardrails for AI agents. Validates shell commands and file reads against 400+ security patterns to block destructive operations, credential theft, and unauthorized system access.