Reference

Lookup material: the command-line surface, how to choose a model, the knowledge-stewardship skill, and a full worked investigation.

• CLI — lore is one static binary: a demo that runs with no key and no cluster, the server, and the knowledge-catalog subcommands.
• Benchmarking — RunLore is model-agnostic, so the model you pick is a real decision. This is how to measure it on your own incidents rather than trusting a leaderboard.
• kb-steward — the skill for seeding, capturing and curating catalog entries. Diagnosis stays RunLore’s job; the skill only writes knowledge down well.
• Examples — a real investigation end to end, from the firing alert to the merged catalog entry.

Looking for the tools the investigation loop calls — query_metrics, what_changed, pod_logs and the rest? Those are described with the backends that unlock them, in Data Sources.