3. Data, Science & AI

Automatically analyzes CSV files using pandas to generate summary statistics and visualizations without user input.

Queries Google NotebookLM for source-grounded, citation-backed answers from documents using Gemini AI, reducing hallucinations with browser automation for document management.

Enables seamless context transfer between AI agents while preserving critical conversation elements and decision flow.

Builds psychometrically accurate AI agents that model user personality and decision heuristics to negotiate and act on their behalf.

Enables text-to-speech conversion using Kokoro TTS AI voice synthesis on Apple Silicon.

Manages car logo data lifecycle including fetching, scraping, storage in Vercel Blob, and cache operations for brand normalization.

Creates and maintains interactive charts for visualizing car registration and COE data, including new chart implementations and bug fixes.

Customizes LLM context for generation, enabling RAG context injection, message history control, and thread-based search.

Enhances AI agents with Retrieval-Augmented Generation (RAG) to search knowledge bases, retrieve context, and ground responses in specific data.

Manages file uploads, image attachments, and media processing for AI agents during conversations, enabling image analysis and document processing.

Offers a no-code web UI for real-time testing, debugging, and behavior verification of AI agents with conversation history browsing capabilities.

Configures Convex agents for chat-based AI interactions, enabling agent initialization, conversation threads, and LLM-generated responses.

Provides a framework for orchestrating multi-agent systems, ensuring context preservation (WHERE, WHAT, WHY) while allowing sub-agents autonomy in implementation (HOW).

Provides local server for GGUF-based LLM inference with OpenAI-compatible API, enabling offline AI tool development and troubleshooting.

Generates Claude Code agents using Anthropic best practices, supporting new agent creation, configuration modification, and specialized subagent design.

Provides SQLite best practices, connection management, and migration system for PolyFlup, ensuring efficient database operations.

Analyzes production database data for trades, balances, and market history to generate actionable financial insights through synchronization and processing.

Automatically compresses context by summarizing completed phases, extending session length by 2-3 times for efficient AI interactions.

An AI-powered memory system enabling users to save, search, and recall knowledge with seamless memU integration for efficient information management.

An AI engine that dynamically generates new capabilities to adapt and enhance user interactions without predefined limitations.

Exports Dev Graph UI assets including SVGs, MP4 timelines, per-commit visuals, and dashboard analytics JSON via API.

Exports UMAP projections and HDBSCAN clusters, and manages Qdrant vector collections for latent space analysis and similarity search.

Analyzes historical data to identify patterns and generate predictive insights using data science techniques.

Deprecated router for database operations, directing to domain-specific database skills (core, workflow, agents, context).