[AZ-446] CSV reporter: band + ci95 annotations + report.csv emitter

Batch 89 — adds optional `band`, `ci95_low`, `ci95_high` kw-only
parameters to `_NfrRecorder.record_metric` and emits a new per-metric
report.csv artifact (one row per scenario × metric, columns:
scenario_id, metric_name, value, value_band, ci95_low, ci95_high,
ac_id, outcome). Backwards compatible — existing 4-arg callers
unchanged; unbalanced ci95 pair raises ValueError. report.csv is
written once per pytest session from `pytest_sessionfinish` so the
annotation pass runs once per CI invocation regardless of
(fc_adapter, vio_strategy) (AC-3). `regression-baseline.json`
intentionally kept flat to preserve the diff contract used by
regression-detection tooling.

NFT-RES-03 + NFT-PERF-01 scenarios updated to pass real bands and
compute empirical 2.5/97.5-percentile ci95 from their own sample
streams (per-iteration envelope ratios for Monte Carlo,
per-frame latency samples for N-sample latency).

Tests: 1229 e2e/_unit_tests pass (+6 vs. batch 88 for AZ-446
band/CI behavior, value-error on unbalanced ci95, report.csv columns,
explicit-path override, and end-to-end emission via the pytest
plugin). Code review: PASS_WITH_WARNINGS — 1 Low (empirical-CI
semantics, documented inline), 1 Medium carried over from batch 88's
cumulative-review backlog (write_csv_evidence + _resolve_fixture_path
duplication is outside AZ-446 reporting scope).

This commit closes Step 10 Implement Tests for cycle 1 (41 of 41
blackbox-test tasks done, AZ-406..AZ-446). Greenfield auto-chains to
Step 11 Run Tests next.

Co-authored-by: Cursor <cursoragent@cursor.com>

_docs/02_tasks/todo/AZ-446_csv_reporter_refinements.md

@@ -1,51 +0,0 @@
-# CSV reporter refinements + per-NFR machine-readable trend lines
-
-**Task**: AZ-446_csv_reporter_refinements
-**Name**: Trend-line + acceptance-band annotations + Monte Carlo confidence intervals
-**Description**: Add trend-line + acceptance-band annotations to the CSV reporter (so each numeric metric carries the AC's threshold inline), and emit Monte Carlo confidence intervals where applicable (NFT-RES-03, NFT-PERF-01).
-**Complexity**: 2 points
-**Dependencies**: AZ-406, AZ-445
-**Component**: Blackbox Tests / Reporting (epic AZ-262)
-**Tracker**: AZ-446
-**Epic**: AZ-262 (E-BBT)
-
-## Problem
-
-A bare numeric metric without its acceptance band is hard to read. Adding the band inline (e.g., `p95_latency_ms = 387 (≤400 budget)`) makes the report immediately legible. Monte Carlo runs additionally need confidence intervals so a single iteration's outlier doesn't drive false-positive failures.
-
-## Outcome
-
-- `report.csv` columns extended: every numeric column gets a paired `_band` column (e.g., `p95_latency_ms`, `p95_latency_ms_band`).
-- Monte Carlo CI: NFT-RES-03 + NFT-PERF-01 emit a `_ci95_low` + `_ci95_high` paired column.
-
-## Scope
-
-### Included
-- Band annotation per numeric column (read from each scenario's task spec).
-- CI-95 columns for the Monte Carlo + N-sample scenarios.
-
-### Excluded
-- Visual report rendering (HTML, etc.) — out of scope.
-
-## Acceptance Criteria
-
-**AC-1: band annotation**
-Given any numeric metric column in `report.csv`
-Then a paired `_band` column exists with the AC threshold text (e.g., `≤400 ms`).
-
-**AC-2: CI-95 for Monte Carlo**
-Given NFT-RES-03 emits 100 iterations
-Then `report.csv` contains `_ci95_low` + `_ci95_high` paired columns for each metric.
-
-**AC-3: parameterization**
-Given conftest parameterization
-Then this annotation pass runs once per CI invocation regardless of `(fc_adapter, vio_strategy)`.
-
-## System Under Test Boundary
-
-Reporting-only.
-
-## Document Dependencies
-
-- `_docs/02_document/tests/test-data.md` § Reporting & Evidence
-- Each scenario task's AC numeric thresholds (read at report time)