Analytics & Monitoring Agent Skills for OpenAI Agents SDK

Track ML experiments, manage model registry with versioning, deploy models to production, and reproduce experiments with MLflow…

Generates code that deploys fine-tuned models from SageMaker Serverless Model Customization to SageMaker endpoints or Bedrock.

When to activate: model evaluation, metrics, AUC-ROC, F1, RMSE, SHAP, calibration, fairness, confusion matrix, bootstrap

Build this skill enables AI assistant to provide interpretability and

Reduce LLM size and accelerate inference using pruning techniques like Wanda and SparseGPT.

MANDATORY tracking protocol for multi-model validation.

Run and analyze molecular dynamics simulations with OpenMM and MDAnalysis.

Molecular featurization for ML (100+ featurizers).

Stream live swarm events using the Monitor tool for real-time observability

Monitoring and observability patterns for Prometheus metrics, Grafana dashboards, Langfuse v4 LLM tracing (as_type, score_current_span…

System health check (MOT) for skills, agents, hooks, and memory

Analyzes 1C:Enterprise Data Composition Schema (SKD) XML files to inspect datasets, query texts, calculated fields, resources, and report…

Operate and troubleshoot the MT5 tick collection system.

This skill should be used when long-running agent sessions need context compression, structured summarization, compaction, token-per-task…

This skill should be used for project-level decisions about LLM-powered systems: whether an LLM is the right primitive for the task at…

When the user wants to set up, improve, or audit analytics tracking and measurement.

Educational GPT implementation in ~300 lines.

Generate structured narrative text visualizations from data using T8 Syntax.

GPU-accelerated data curation for LLM training.

Guide developers through capturing diagnostic artifacts to diagnose production .NET performance issues.

Instrument API requests with spans and distributed tracing.