[AZ-492] Cycle 3 batch 4: perf harness PT-07 + PT-08 + JWT-attach

Drains all three deferred perf-harness items in one batch:
- PT-01..PT-06 now carry Authorization: Bearer minted via the canonical
  SatelliteProvider.TestSupport.JwtTokenFactory (AZ-491) — no third copy
  of JWT logic in the shell.
- PT-07 implemented as cold + warm dual-pass distribution (N=20 each),
  reports p50/p95 for both passes and fails if warm p95 >= cold p95.
- PT-08 implemented as 20-batch upload distribution with batch p95 gated
  at the AZ-488 2000 ms target; per-item gate cost reported as derived
  proxy (batch_p95 / batch_size).

New SatelliteProvider.IntegrationTests/PerfBootstrap.cs adds two CLI
short-circuit subcommands (--mint-only and --gen-uav-fixture <path>)
invoked by the shell so the perf script never inlines the JWT or
JPEG-fixture logic. The dispatch sits at the top of Program.cs Main
and runs before any HTTP / DB / readiness setup.

performance-tests.md PT-07 + PT-08 flip from Deferred to Implemented.
traceability-matrix.md PT-07 + PT-08 rows move from recorded to covered
(PT-08 partial due to per-item proxy — flagged Low in batch-4 review).
_docs/_process_leftovers/2026-05-11_perf-pt07-harness.md deleted; the
leftovers directory is now empty.

Closes cycle-2 retro Action 2; LESSONS.md [process] rule about Deferred
NFRs remains in force as a guardrail.

Also includes the previously-uncommitted cumulative review report for
cycle-3 batches 01-03 (generated at the end of batch 3 but not staged).

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

_docs/06_metrics/retro_2026-05-11_cycle2.md

@@ -137,7 +137,7 @@ This is NOT a recommendation to never accept user-overrides on SP cap — the cy
 **Owner**: whoever picks up the next test-infrastructure PBI. Recommend scheduling as `AZ-XXX Unify test JWT mint helper` at 3 SP.
 **Estimated impact**: removes the dual-fix tax on every future JWT-side change; prevents drift on the surface that mints test credentials.
 
-### Action 2 — Schedule PT-07 harness work as actual feature work, not a leftover
+### Action 2 — Schedule PT-07 harness work as actual feature work, not a leftover — **RESOLVED in cycle 3 (AZ-492)**
 
 **Why**: Pattern 3 + Comparison checkpoint. Cycle 1 Action 2 made the spec-side honest (no more silent Unverified scenarios) but did nothing to drain the backlog. Cycle 2 added a 3rd deferred item (the script-rot on `run-performance-tests.sh`). The perf gate has been 0-of-N for two cycles running and is now actively misleading — anyone reading the gate banner sees "skipped, will check next cycle" without realizing "next cycle" never arrives.
 
@@ -149,6 +149,8 @@ This is NOT a recommendation to never accept user-overrides on SP cap — the cy
 **Owner**: next planning loop. Sequence it before any AC that has a hard latency or throughput requirement.
 **Estimated impact**: turns the perf gate from theatre into a real gate. Reduces the surface area where regressions can hide.
 
+**Cycle 3 resolution (AZ-492)**: PT-07 + PT-08 + Bearer-token attach for PT-01..PT-06 all landed in AZ-492 (cycle 3 batch 4). Leftover `_docs/_process_leftovers/2026-05-11_perf-pt07-harness.md` deleted. PT-07 and PT-08 status flipped from `Deferred` to `Implemented` in `_docs/02_document/tests/performance-tests.md`; traceability-matrix.md PT-07 + PT-08 rows moved from `◐ recorded` to `✓`. The cycle-1 `[process]` LESSONS.md rule about Deferred-status NFRs remains in force as a guardrail.
+
 ### Action 3 — Reset integration-test DB state between runs (real fix, not workaround)
 
 **Why**: Pattern 5. The wallclock-seed for `_coordinateCounter` is a workaround that lets cycle 2 ship but doesn't fix the fundamental issue: integration tests are not isolated from the persistent Postgres volume. The next test class that inserts into `tiles` will have to invent its own collision-avoidance scheme. Eventually two tests will collide despite both using wallclock seeds (e.g. parallel test execution, fast back-to-back runs in CI).