Lesson 053: Eval harness hardening
Focus
Anchor this drill to one production LLM workflow—even hypothetical. Token Eval harness hardening:53 keeps neighbouring lessons differentiable.
Key ideas
- Thread: Eval harness hardening · drill v3 · spin
689874. - Habit: attach a trace_id to every completion you would paste into an ops dashboard.
- Guardrail: add one RACI bullet for prompt or index changes before tomorrow's standup.
Deep dive notebook
Synthetic drill artefacts
Tool contract rehearsal
{
"name": "reindexSlice",
"arguments": { "corpus": "corpus-5", "dry_run": true }
}
Preconditions: index lag < 14h
FAILURE: page platform if lag > SLA without VP memo.
Practice
Practice Pair with security on one prompt-injection tabletop scenario. — 53 Bump 15.