3. Data, Science & AI

Creates interactive, responsive data visualizations including dashboards, network graphs, and geospatial maps using D3.js.

Comprehensive tool for creating, editing, and analyzing Excel spreadsheets with formulas, charts, and data transformation capabilities for data-driven insights.

Guides creation of Model Context Protocol (MCP) servers for integrating external APIs, tools, and authentication with Claude AI agents.

Analyzes agent reasoning loops, step functions, and termination conditions to understand and document AI agent framework implementations.

Analyzes tool registration, schema generation, and error feedback in agent frameworks to evaluate interface design and LLM interaction mechanisms.

Analyzes AI agent frameworks to inform system design, compare best practices, and document architectural decisions with evidence.

Analyzes context management, memory systems, and state continuity in LLM agent frameworks, including prompt assembly, eviction policies, and token budgeting.

Analyzes coordination patterns, handoff mechanisms, and state sharing in multi-agent AI systems to evaluate communication protocols and orchestration approaches.

Adds and manages evaluation results in Hugging Face model repositories using .eval_results/ format, extracting scores from model cards.

Analyzes protocol layer between agent harnesses and LLM models, covering message formats, tool calls, streaming, and agentic primitives.

Quick reference for LangChain framework concepts including agents, RAG, vector stores, and building LLM applications.

Intelligent web scraper for extracting content from websites with multiple output formats and robust error handling.

Fetches Estonian statistical datasets and metadata from the Statistics Estonia API, including ESMS methodology and quality documentation for time series analysis.

Provides interaction with a Text Processing AI Agent for text analysis, formatting, and processing tasks.

Fetches LLM-optimized documentation for Supabase, enabling seamless integration of Supabase features into AI applications and RAG systems.

Packages codebases into context-optimized formats for efficient AI analysis and comprehension.

Mitigates context failures including lost-in-middle and poisoning to enhance AI model reliability and response accuracy.

Calculates a SCUM score metric for data quality assessment and AI model performance evaluation.

Optimizes AI model context window using compression, masking, caching, and partitioning techniques.

Provides foundational knowledge on context engineering for AI agents, covering context windows, attention mechanics, and context budgeting strategies.

Designs effective tools for AI agents, covering descriptions, consolidation, architecture, naming conventions, and error handling.

Fetches up-to-date Expo documentation using LLM-based retrieval for React Native development references.

Processes SARIF files from static analysis tools, enabling aggregation, deduplication, and filtering of analysis results.

Guides creation of custom subagents for Claude Code, including system prompts, skill selection, and agent file generation.