Skill UI

Provides patterns and techniques for building robust evaluation loops for AI agents. This skill guides the implementation of self-critique, evaluator-optimizer pipelines, and test-driven refinement workflows. Use it to significantly enhance the accuracy and quality of complex outputs like code, reports, and analyses by implementing structured, iterative improvement cycles (Generate → Evaluate → Critique → Refine).

Provides comprehensive patterns and techniques for securing AI agent systems. Learn how to enforce governance policies, including tool allow/block lists, content filtering, rate limiting, and human approval workflows. Essential for building robust, compliant, and trustworthy agents in multi-agent or production environments that interact with external APIs and sensitive data.

A comprehensive meta-skill template designed for developers to scaffold, structure, and define specialized Agent Skills for AI platforms like GitHub Copilot. It guides users through mandatory requirements, including folder naming conventions, YAML frontmatter best practices (name, description), and structuring the skill body with detailed workflows and documentation sections. Use this when building complex, specialized AI capabilities.

Connects to Google NotebookLM to query your uploaded documentation via Gemini’s source-grounded answers, automating authentication, notebook management, and question workflows through the provided run.py wrapper.

The Phoenix CLI is a powerful tool for debugging complex LLM and agent workflows. It allows users to fetch detailed traces, spans, and sessions, analyze errors, and structure findings using open and axial coding techniques. It is essential for investigating complex LLM failures, building robust failure taxonomies, and guiding the next steps in application evaluation after initial instrumentation.

Phoenix Evals lets teams build evaluators for AI/LLM apps by wiring deterministic code with LLM nuance, running Python or TypeScript workloads, validating against humans, and iterating through error analysis, RAG, and production workflows.

Guides hooking Phoenix AI services with OpenInference tracing, covering setup, spans, instrumentation, and production deployment for LLM workflows using Python or TypeScript.

Consult another agent for help, plans, reviews, or decision support in multi-agent workflows. Explains using list_agents() and chat_with_agent(), background submission and status checks, session ID reuse, and loop-avoidance best practices.

Use the esm Python SDK with ESM3 and ESMC for protein sequence generation, embeddings, function conditioning, structure prediction, inverse folding, and Forge/Biohub inference workflows.

Plans, audits, and safely runs ChicagoHAI HypoGeniC/HypoRefine workflows for LLM-assisted hypothesis generation from labeled text datasets, with pinned installs, config validation, dataset leakage checks, and bounded local review before any model call.

PyTDC enables discovery of therapeutic machine learning tasks, access to approved datasets, task-aware data splits, evaluation metrics, benchmark groups, and molecular oracle workflows. Designed for biomedical research and drug discovery applications using Python.

PyTorch Lightning is a deep learning framework that organizes PyTorch code to eliminate boilerplate while maintaining flexibility. It automates training workflows, supports multi-GPU/TPU orchestration, and implements best practices for neural network training. Features include LightningModules, Trainers, data pipelines, callbacks, logging (W&B, TensorBoard), and distributed training strategies like DDP and FSDP for scalable training.