Define the task
Share the goal, source material, constraints and what a useful answer should include.
Machine learning and LLM evaluation.
Frame ML problems, inspect datasets, design reproducible training, assess model quality and evaluate retrieval-augmented generation.
01 / HOW TO USE GNCORETASK / SKILL / EVIDENCE
Pick a skill
Choose a focused skill ID from the domain, then install it for the appropriate assistant.
Review the output
Check the result and evidence against your task. The assistant—not GNCore—carries out any execution.
02 / RELATED REAL SKILLSFROM GNCORE 4.0.0
/ml-dataset-audit
ML Dataset Audit
Audit dataset provenance, schema, representativeness, label quality, leakage, and split strategy.
Skill details & install ↗/ml-llm-evaluationLLM Evaluation Suite
Evaluate an LLM feature with representative tasks, rubrics, adversarial cases, latency, and cost.
Skill details & install ↗/ml-ragRAG Grounding Workflow
Build and evaluate retrieval-augmented answers with source access controls and abstention.
Skill details & install ↗03 / TRY A WORKFLOWCOPY / INSTALL / INVOKE
Start focused.
Stay in control.
Training, model access and evaluation tools must come from the assistant environment.
CLI / EXAMPLE
gncore search "machine"
gncore install ml-dataset-audit --apps vscode-codex --scope workspace