Task — engineering-spec@1

"Golden recall eval runner + 25 seed cases wired into CI"

draftTASK-MEMORY-209

module memory · class product · priority p1 · created 2026-07-08 · shipped null
depends on none · blocks none

TASK-MEMORY-209: Golden recall eval runner + 25 seed cases wired into CI

1. Description

memory-eval binary scores recall@10 and MRR against seed set; CI gate fails on >3% drop

Migrated 2026-07-08 from the memory improvement backlog, folded into the task system as class: improvement. Source report refs: R45, F32.

Acceptance criteria

[ ] memory-eval binary scores recall@10 and MRR against seed set; CI gate fails on >3% drop