Skill
GenAI evaluation & safety
Hallucinations, jailbreaks, and eval harnesses. If you cannot measure it, you cannot improve it.
Do these first
What you will be able to do
- Build a 20-example eval set
- Name prompt injection as a security issue
- Refuse disallowed tasks in the system spec
Curated resources
Start with the pick, then branch out
Scores are LearnPath editorial opinions (clarity, cost, freshness, project density) — not an objective ranking. Every YouTube item is embedded officially after a click-to-load facade.
Official docs
LearnPath pickdocumentationintermediate
OWASP Top 10 for LLM Applications
OWASP50 min
Prompt injection is a security bug, not a vibe.
Score 91
Open
55 minLearnPath pickMore curatedyoutube courseintermediate
How to Systematically Setup LLM Evals (Metrics, Unit Tests, LLM-as-a-Judge)
Dave Ebbelaar55 min
Fifty-five minutes that treat evaluation as engineering: define metrics, write unit tests for prompts, use LLM-as-judge carefully, and track regressions. The most actionable evaluation resource for people shipping LLM features.
Score 82
Watch