[Docs] (.claude/skills, .github/workflows, CMakeLists.txt, docs/archive, docs/Disaggregated, docs/MobileVA, docs/Performance, MG_Backend, MG_Backend/DirectGLES, MG_Backend/DirectVulkan, MG_Backend/Record, MG_Gbm, MG_Impl, MG_Impl/EGLImpl, MG_Impl/GLImpl, MG_Impl/GLXImpl, MG_Impl/Pipe, MG_IntegrationTest, MG_IntegrationTest/Harness, MG_IntegrationTest/Scenarios, MG_Pipe, MG_Pipe/generated, MG_Remote, MG_Remote/Client, MG_Remote/Server, MG_Remote/Transport, MG_Remote/Wire, MG_State/GLState, MG_Test/Pipe, MG_Test/Util, MG_Test/Wire, MobileGL, scripts, scripts/data, tools/device_bench, userguide): archive the planning notes and agent contracts under docs/archive

- DEBTS.md and OPEN-QUESTIONS.md stay maintained, now in docs/Disaggregated.

tools/device_bench/disagg/README.md

@@ -82,7 +82,7 @@ FCL's Minecraft renders into the Render Server's on-screen window on the same ph
 `fcl_accept.sh <backend> kill` kills the server and checks the client latches DEVICE LOST and FCL
 survives. Needs the FCL fork's version setting "keepRunningInBackground" ON in `config.json`
 (FCL pauses the game the moment its Activity leaves the foreground). The server is started through
-the Render Server screen's documented extras. Results: `docs/Disaggregated/notes/p12/FCL-ACCEPTANCE.md`.
+the Render Server screen's documented extras. Results: `docs/archive/notes/p12/FCL-ACCEPTANCE.md`.
 
 ## crosshost_accept.sh + crosshost_report.py — P12 exit gate (b)
 
@@ -90,7 +90,7 @@ A WSL client replays the openra and rd12 traces against the phone's render serve
 TCP + stream (ssim gate, then `PRESENT_CREDIT` 1/2/3 benchmarks). Adds a temporary /32 route
 around the host's TUN proxy, checks the path with `tcp_path_check.sh`, and removes the route and
 stops the server on exit. `crosshost_report.py` reduces the output to a table. Results and the
-P6.5 link numbers: `docs/Disaggregated/notes/p12/CROSSHOST-ACCEPTANCE.md`.
+P6.5 link numbers: `docs/archive/notes/p12/CROSSHOST-ACCEPTANCE.md`.
 
 ## netbench.sh — Wi-Fi TCP throughput baseline
 
@@ -133,7 +133,7 @@ Reducers, all reading a probe directory:
 - `fcl_stacks.py`: top stacks.
 
 The simpleperf reducers need `SIMPLEPERF_DIR` and a `binary_cache` from `binary_cache_builder.py`.
-Findings: `docs/Disaggregated/notes/perf-p15/PLAN-P15.md` section 2.
+Findings: `docs/archive/notes/perf-p15/PLAN-P15.md` section 2.
 
 ## Clocks, bench sessions and run validity (P15)
 

tools/device_bench/p6/ab_session.py

@@ -13,7 +13,7 @@ Its first use was `CONTRACT-P6.md` section 9 item 7 (the paired device gate) and
 doorbell ledger), and it is not specific to those: the thing it computes is "N transports, one
 thermal window, identical protocol" for any transport/backend/case triple the runner accepts.
 Everything it does is a step of the protocol in docs/Disaggregated/MEASUREMENTS.md sections 4.4 / 9
-/ 11 plus notes/p5f/device-report.md.
+/ 11 plus docs/archive/notes/p5f/device-report.md.
 
 Why it is a script and not a shell transcript: a run that wants to be quotable needs the CPU
 samples, the pin evidence and the library's own wait ledger to come from the SAME window, and the
@@ -26,7 +26,7 @@ WHAT IT DELIBERATELY DOES NOT DO.
 - It does not decide anything about correctness. Benchmark mode takes no snapshot and compares no
   golden (--benchmark replays the whole trace for timing), so an SSIM column here would be a
   fabrication. Correctness on this device is section 3 of the P6 closing snapshot
-  (docs/Disaggregated/notes/p6/README.md) and the retrace lanes.
+  (docs/archive/notes/p6/README.md) and the retrace lanes.
 - It does not read `MOBILEGL_TRANSPORT` out of `--transport` on the runner. That flag is applied by
   run_android_retrace_local.py's run_case() path only; run_benchmark_case() passes `args.env`
   alone, so `--transport spawn` alongside `--benchmark` is silently ignored and the arm runs
@@ -821,7 +821,7 @@ def main():
 
     # The fan is a CONSTANT of the protocol, not a treatment: both arms run under the same airflow,
     # and on this device it is the difference between a 60 s cool-down between arms and a 20-minute
-    # one (notes/p4a/INTEGRATOR-DECISIONS.md). `real_speed` is read back because target_level alone
+    # one (docs/archive/notes/p4a/INTEGRATOR-DECISIONS.md). `real_speed` is read back because target_level alone
     # is a request, not a fact.
     fan = "/sys/class/xm_power/hw_monitor/pwm_fan"
     adb.su(f"echo {args.fan_level} > {fan}/target_level")