3. Data, Science & AI

Automates Excel spreadsheet operations including data import, export, report generation, and workbook manipulation using popular libraries.

Detects declining AI output quality and automatically re-anchors prompts to resolve context confusion or user feedback indicating issues.

Automates conversation context management by clearing outdated data while preserving key knowledge via context markers, triggered by user commands or token limits.

Manages user preferences and corrections for bilateral modeling, auto-learning from session feedback to adapt to user choices.

Displays session efficiency report with token savings, cache performance, and optimization recommendations for Navigator usage.

Retrieves cross-session memories of past decisions and interactions to provide context-aware responses and conversation continuity.

Guides users on utilizing TB2J command-line tools to compute magnetic interaction parameters (J, DMI) from DFT outputs (Wannier90, Siesta, Abacus).

Ingests procedural knowledge (skills, patterns, documents) into vector databases for AI-driven retrieval and processing.

Initializes and verifies configuration of an AI agent's memory system, ensuring proper setup for subsequent operations.

Provides a guide for integrating DSPy with Microsoft Agent Framework in AgenticFleet, covering typed signatures, assertions, and agent handoffs.

Provides structured initialization for AI agent memory and project context, enabling effective codebase learning and workflow configuration.

Enables semantic search over stored memory contexts using Chroma Cloud to retrieve solutions, patterns, and context for AI-driven applications.

Quick reference for DSPy agent framework integration patterns including typed signatures, assertions, routing cache, and agent handoffs.

Enables developers to build data-driven features and integrate AI/ML capabilities through data analytics and machine learning expertise.

Guides users on compiling AI agent skills from live sources using the Colin framework for skill management and project development.

Creates and manages example implementations for AI SDK functions to validate provider support, demonstrate features, and serve as test fixtures.

Framework enabling AI agents to make informed technical decisions on architecture, implementation, and trade-offs, enhancing agent-driven system design.

Guides creation and enhancement of Claude Code agents, including structure, persona, tool access, model selection, and schema validation.

Exports matrix data, memory, solutions, failures, and warnings to portable formats for backup or transfer.

Fetches and packages code from external repositories (e.g., GitHub) for context analysis in AI systems, such as RAG.

Extracts audio from Douyin/TikTok videos and transcribes it to text with precise timestamps for easy reference.

Normalizes JSONL interfaces across skills to ensure consistent fields (ids, titles, citation keys) for deterministic downstream processing without best-effort joins.

Defines evidence-based table schemas for surveys, ensuring each column maps to verifiable evidence sources without placeholders.

Retrieves arXiv paper metadata via keyword search and saves results in structured JSONL format for research datasets.