Milestone 3.5.F.3 — Benchmark Sign-Off and ADR Index

Status: Planned
Goal: Track F — Exit Gate
Phase: 3.5 — Extraction & Context Assembly
Estimated effort: 1 day
Track: Track F — Exit Gate
Depends on: Milestones 3.5.F.1–3.5.F.2

Why This Milestone Exists

Extend benchmarks, publish ADRs, and write the Phase 3.5 exit-audit doc following the established pattern.

Deliverables

Target outcomes for the milestone; concrete artifacts may differ from any sketch above.

Extend benchmarks/README.md profiles to cover: proxy overhead with context assembly, MCP tool-call p95, extraction-worker throughput
Publish ADR-0038 (context assembly — move from Pending to Accepted); cite ADR-0050 for MCP transport/auth (do not reuse ADR-0043/0044 numbers — those are tokenizer / response-pipeline)
Write decisions.mdx (Phase 3.5 decision log) and phase-3.5-exit-audit.md (P0–P3 gap tracking)

Phase 3.5 Exit Criteria

Extraction worker converts a completed session into >=1 correctly-categorized memory
Subsequent chat request transparently injects a relevant memory
Zero cross-tenant memory leakage across all ISO-* and ISO-MCP-* cases
Context assembly p95 < 50ms at 20 concurrent requests; proxy overhead <20ms p99
All 4 degradation-ladder levels independently verified
MCP tools function end-to-end, sharing one substrate with proxy path
MCP server fails closed on auth-service outage; independent rate-limit budget verified
All Phase 1, 2, and 3 regression tests pass with zero regressions
make e2e-smoke-p3.5 exits 0
ADR-0038, ADR-0043, ADR-0044 published and indexed

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
MCP server surface

Success signals

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

Benchmark profiles extended and baselines committed
ADRs published in web/content/docs/adr/meta.json
Phase 3.5 decision log and exit-audit doc committed
All exit criteria verified
Repo guards / CI checks still pass

Prerequisites

Milestones 3.5.F.1–3.5.F.2