Milestone 5.D.1 — Retrieval Quality Sign-off

Status: Planned
Goal: Track D — Exit Gate
Phase: 5 — Advanced Retrieval
Estimated effort: 2 days
Track: Track D — Exit Gate

Exit Criteria

Hybrid retrieval (BM25 + RRF) live, default-on, beating dense-only on gold-set v2 lexical queries
Cross-encoder reranker meets latency budget within context-assembly p95 ≤ 50ms / p99 ≤ 100ms
Graph traversal live via API and MCP tool, with ISO-MCP-LINEAGE-01 passing
Gold-set v2 committed with inter-annotator agreement ≥0.8 kappa
Multi-mode benchmark running in CI on retrieval-touching PRs
Shadow-evaluation pipeline live with working divergence dashboard
All Phase 1–4.5 regression suites green
ADR index updated
make e2e-smoke-p5 exits 0
Phase 5 gap register published with zero open P0s

Orientation (indicative)

Named paths, package layouts, libraries, schemas, env vars, and commands anywhere on this page are rough sketches for orientation — inspiration and a baseline, not a required change list.

During implementation, expect to:

open the live tree and follow existing patterns before inventing new ones
research current constraints (latency, tenancy, deploy shape, libraries) more deeply than this page can
advance the design beyond the sketch where measurement or code reality says so
land work in different filenames, merged packages, deferred docs, or new surfaces when the situation calls for it

Prefer outcomes over matching any particular file tree or command sequence.

Areas that may be involved (situational — not a checklist):

Evaluation / benchmarks

Success signals

Outcome-oriented signals that the milestone is in good shape. Exact filenames, package layouts, and commands may differ from any sketches above.

All 10 exit criteria verified
Regression gate passes in CI
No unaddressed benchmark regressions

Prerequisites

All Track A–C milestones complete