feat(simulation): qualify D-only stock rover smoke

docs/12_SIMULATION_POLYGON_PRODUCT_AND_SRS.md

@@ -51,11 +51,14 @@ accepted navigation behavior, safety behavior or real actuator control.
 
 Polygon is now a parallel product branch. As of this document:
 
-- P0 architecture and SRS are being committed;
+- P0 architecture and SRS are committed on the Polygon branch;
 - SIM S0 has a strict candidate qualification profile and read-only doctor;
-- no target-worker S0 evidence has been admitted;
-- PX4/Gazebo/ROS 2/Nav2 have not been installed or accepted by this repository;
-- no stock rover run has been accepted;
+- the dedicated D-only worker distribution has been created and the candidate
+  PX4/Gazebo/ROS 2/Nav2 stack has been installed without using Docker;
+- PX4 v1.17.0 and the stock Ackermann rover have completed repeated factual
+  smoke runs with ROS 2 vehicle status, Gazebo clock and clean shutdown;
+- the complete digest-bound S0 evidence manifest has not been admitted;
+- pause/step/speed and accepted 1×/2× resource evidence remain open;
 - `actuator_authority=false`;
 - `navigation_or_safety_accepted=false`.
 
@@ -369,7 +372,8 @@ incompatible.
 The S0 candidate line is:
 
 - Windows 11 AI worker;
-- WSL2 Ubuntu 24.04;
+- dedicated `MissionCore-Sim` WSL2 Ubuntu 24.04 distribution physically on D;
+- no Docker dependency in S0, so active shared Docker co-tenants remain untouched;
 - ROS 2 Jazzy;
 - Gazebo Harmonic;
 - PX4 Autopilot v1.17.0;
@@ -378,9 +382,10 @@ The S0 candidate line is:
 - Nav2 Jazzy;
 - PX4 Gazebo stock rover models.
 
-The exact source commits already resolved in the candidate profile do not become
-accepted merely because they are present. ROS/Gazebo package versions and every
-candidate build remain pending target-worker qualification.
+The exact source commits and package versions are resolved in the candidate
+profile. Target installation, PX4 build, DDS telemetry and repeated stock-rover
+smoke evidence now exist, but components do not become `accepted` until the
+remaining S0 behavior/resource cases and digest-bound evidence generation pass.
 
 S1 uses direct exact `px4_msgs` integration. The experimental
 `px4-ros2-interface-lib` is evaluated later and is not on the S1 critical path.
@@ -390,8 +395,9 @@ S1 uses direct exact `px4_msgs` integration. The experimental
 All mutable worker bytes must be physically placed under
 `D:\NDC_MISSIONCORE`, including:
 
-- WSL distribution/VHD;
-- Docker data and images;
+- the dedicated `MissionCore-Sim` WSL distribution/VHD;
+- Docker data and images if a later profile elects to use containers; the S0
+  profile explicitly does not;
 - sources and build products;
 - package/model caches;
 - runtime and logs;
@@ -402,6 +408,13 @@ The WSL view is `/mnt/d/NDC_MISSIONCORE`. C-drive placement is a blocking
 failure. S0 requires at least 300 GiB free before installation and a stop-below
 threshold of 200 GiB. Full external datasets are out of S0 scope.
 
+S0 runtime processes execute in a new loopback-only Linux network namespace.
+Micro XRCE-DDS Agent v2.4.3 has no address-selection flag and therefore binds
+UDP wildcard inside that namespace; the namespace contains only `lo`, so the
+effective exposure remains loopback-only. ROS domain 0 is isolated by the same
+boundary. This distinction is part of the port evidence and must not be
+misreported as a host-wide `127.0.0.1` bind.
+
 ## 16. Delivery plan
 
 | Phase | Estimate | Exit condition |
@@ -435,7 +448,10 @@ Implementation cards follow this order:
 SIM S0 requires all of the following factual evidence:
 
 - worker inventory;
-- physical D-only placement, including WSL and Docker VHD;
+- physical D-only placement of the dedicated WSL VHD, sources, builds, caches
+  and artifacts;
+- confirmation that S0 uses the dedicated D-only WSL and does not use the
+  existing C-backed Docker Desktop runtime;
 - exact component/package pins and licenses;
 - one stock PX4 rover launch;
 - ROS 2 telemetry;
@@ -444,7 +460,7 @@ SIM S0 requires all of the following factual evidence:
 - clean stop with no orphan processes;
 - 1× resource baseline;
 - 2× resource baseline;
-- explicit `GO`, `PAUSE` or `BLOCKED` report.
+- explicit `GO`, `INCOMPLETE` or `BLOCKED` report.
 
 The repository profile and doctor implement preflight and evidence admission.
 They do not install packages, start processes, open ports or write a target

docs/adr/0015-simulation-polygon-qualification-boundary.md

@@ -44,13 +44,20 @@ reports easy to misuse.
 9. Source-of-record, derived artifacts and presentation state remain separate.
    Reset creates a new immutable run/episode.
 10. Ground truth is evaluator-only and is not visible to perception or planning.
-11. SIM S0 targets the reviewed Windows 11/WSL2 AI worker with all mutable bytes
-    physically under `D:\NDC_MISSIONCORE`.
-12. S1 uses exact `px4_msgs` directly. The experimental
+11. SIM S0 targets the reviewed Windows 11 worker through a dedicated
+    `MissionCore-Sim` WSL2 Ubuntu 24.04 distribution physically under
+    `D:\NDC_MISSIONCORE`.
+12. SIM S0 does not depend on the existing Docker Desktop runtime. Its WSL
+    backend is C-backed and currently serves unrelated co-tenants.
+13. S1 uses exact `px4_msgs` directly. The experimental
     `px4-ros2-interface-lib` is not a mandatory S1 dependency.
-13. A top-level Polygon UI is gated on accepted contracts and a proven S1
+14. A top-level Polygon UI is gated on accepted contracts and a proven S1
     backend lifecycle with persisted run history.
-14. Real actuator authority requires a new decision and physical safety gate.
+15. Real actuator authority requires a new decision and physical safety gate.
+16. S0 processes run inside an ephemeral loopback-only Linux network namespace.
+    Micro XRCE-DDS Agent's UDP wildcard bind is acceptable only inside that
+    boundary; no S0 listener may be reachable from the worker LAN or Windows
+    host network.
 
 ## Consequences
 
@@ -67,7 +74,7 @@ reports easy to misuse.
 
 ## First implementation
 
-The first implementation is deliberately read-only:
+The qualification doctor is deliberately read-only:
 
 - `simulation/s0/qualification-profile.yaml` is the candidate stack,
   storage, clock, authority, port and process contract.
@@ -77,6 +84,12 @@ The first implementation is deliberately read-only:
   starting or modifying anything.
 - MISSIONCOR-40 tracks factual worker qualification.
 
+The first target bootstrap also includes `simulation/s0/worker-smoke.sh`. It
+creates the reviewed loopback-only namespace, starts only the pinned stock-rover
+providers, observes ROS 2 vehicle status and Gazebo clock, requests PX4 shutdown,
+checks for owned-process residue and writes D-only factual artifacts. It never
+arms the vehicle or sends an actuator command.
+
 The doctor reports `INCOMPLETE` until it is running on the reviewed target with
 resolved/accepted pins and all required evidence. It cannot infer `GO` from the
 repository state.

docs/lab/006_SIM_S0_WORKER_BOOTSTRAP_20260724.redacted.md

@@ -0,0 +1,126 @@
+# Lab 006: SIM S0 worker bootstrap and stock-rover smoke
+
+Date: 2026-07-24 (Europe/Moscow).
+
+## Scope
+
+The reviewed Windows 11 AI worker received a dedicated Docker-free
+`MissionCore-Sim` WSL2 Ubuntu 24.04 distribution physically stored at
+`D:\NDC_MISSIONCORE\simulation\wsl\MissionCore-Sim`. Source, build, cache,
+runtime, artifacts and datasets are all under the same D-only root. Existing
+Docker Desktop co-tenants were inventoried but not stopped, reconfigured or
+used.
+
+This lab grants no real or simulated actuator authority. Every smoke run kept
+`actuator_authority=false`; it observed the stock rover while disarmed.
+
+## Redacted worker inventory
+
+- Windows 11 Pro, build 26200;
+- WSL2 kernel `6.18.33.1-microsoft-standard-WSL2`;
+- Ubuntu 24.04.4 LTS, x86_64;
+- 13th Gen Intel Core i9-13900KF, 32 WSL-visible logical CPUs;
+- 50,427,760,640 WSL-visible RAM bytes and 12,884,901,888 swap bytes;
+- NVIDIA GeForce RTX 4090, driver 610.47, 24,564 MiB VRAM;
+- D capacity 2,000,381,014,016 bytes;
+- D free after bootstrap and builds: 392,758,685,696 bytes.
+
+The remaining D capacity is above the 300 GiB S0 start threshold and 200 GiB
+stop threshold. This does not waive continuous disk monitoring.
+
+## Installed candidate generation
+
+- ROS 2 Jazzy `ros-base=0.11.0-1noble.20260616.084325`;
+- `ros-gz=1.0.22-1noble.20260616.074726`;
+- Gazebo Harmonic `1.0.0-1~noble`, Gazebo Sim `8.14.0-1~noble`,
+  SDFormat `14.9.0-1~noble`;
+- PX4 Autopilot v1.17.0 commit
+  `d6f12ad1c4f70ad3230afd7d86e971421e02fef4`;
+- `px4_msgs` v1.17.0 commit
+  `86d8239e962f6939e05c3737784f60c02fa884db`;
+- PX4 Gazebo models commit
+  `b6127f4ec20de867e215fb5f78ae88b80f371909`;
+- Micro XRCE-DDS Agent v2.4.3 commit
+  `73622810d984349b80bbac0ef55fc0b694d62222`;
+- Nav2 `navigation2=1.3.12-1noble.20260615.181551` and
+  `nav2-bringup=1.3.12-1noble.20260616.082701`.
+
+PX4 SITL and `px4_msgs` were built successfully from the pinned D-only source.
+The Micro XRCE-DDS Agent was built from source and installed under the D-only
+runtime root. Nav2 is installed but has not yet passed an S2 navigation case.
+
+## Observed smoke result
+
+Two consecutive canonical runs and one preliminary resource run passed:
+
+- `20260724T134400Z-stock-rover`;
+- `20260724T134600Z-stock-rover-repeat`;
+- `20260724T135000Z-stock-rover-baseline`.
+
+The final smoke-script generation also passed as
+`20260724T135800Z-stock-rover-final` and added a runtime listener inventory plus
+the point resource snapshot.
+
+Each run started PX4 v1.17.0 with Gazebo world `rover` and model
+`rover_ackermann_0`, established a Micro XRCE-DDS session, observed
+`/fmu/out/vehicle_status_v1`, observed Gazebo `/clock`, requested PX4 shutdown
+and found no owned runtime residue. The captured vehicle state was disarmed,
+`failsafe=false`, with pre-flight checks passing.
+
+Each provider set ran inside a new Linux network namespace containing only
+`127.0.0.1` and `::1`. The Agent's UDP `0.0.0.0:8888` bind was therefore
+effective only inside that loopback boundary. No PX4, Gazebo or Agent service
+was exposed to the LAN.
+
+The two canonical artifact directories are 68 KiB each. The preliminary
+resource run completed in 38.75 seconds. One point sample observed approximately
+202 MiB RSS for headless Gazebo, 19 MiB for PX4 and 17 MiB for the Agent. The
+worker GPU was already occupied by unrelated co-tenants; no S0 GPU allocation
+was attributable in this headless run. These point observations are not the
+accepted 1× or 2× resource baselines.
+
+## Diagnostics retained
+
+Earlier bootstrap attempts remain outside Git under the D-only artifact root:
+
+- one check raced the asynchronous DDS connection even though the session
+  subsequently established;
+- one used an unsupported ROS 2 Jazzy long option;
+- one used ROS domain 42 while PX4 v1.17.0 published on domain 0.
+
+The smoke contract now waits for the Agent session plus the PX4
+`vehicle_status_v1` writer, uses Jazzy `topic list -t`, and relies on the
+loopback-only namespace to isolate domain 0.
+
+An initial malformed interactive log repeatedly captured PX4 terminal repaint
+output and reached approximately 1.1 GiB. That generated diagnostic artifact
+was removed, then the case was rerun non-interactively; the valid standalone
+PX4 log was approximately 8 KiB.
+
+## Target doctor
+
+The exact candidate profile was executed through the read-only doctor inside
+`MissionCore-Sim`. Profile SHA-256:
+`70bf0df44c85dd57eb3cedb98a474a46ba498c6ffc60ce0d8a06829e28587df1`.
+
+Target identity, authority, clock, port/process registries, all seven D-only
+mutable roots, the disk threshold and all eight required executables passed.
+Version acceptance and the nine-item digest-bound evidence set remain
+incomplete, so the doctor correctly returned `INCOMPLETE`.
+
+## Verdict
+
+`INCOMPLETE`.
+
+Worker compatibility, D-only placement, exact candidate pins, PX4 build,
+stock-rover launch, DDS telemetry, authoritative clock observation and repeated
+clean lifecycle are factually proven. S0 still requires:
+
+- pause/resume and deterministic single-step evidence;
+- 1× speed/real-time-factor acceptance;
+- isolated 2× resource/coexistence baseline;
+- a complete digest-bound evidence manifest;
+- final `GO`, `INCOMPLETE` or `BLOCKED` decision.
+
+This lab does not accept Nav2 behavior, navigation safety, real-time guarantees,
+or real actuator control.

docs/runbooks/SIM_S0_AI_WORKER.md

@@ -5,11 +5,17 @@ real vehicle control or a `GO` verdict by itself.
 
 ## Safety and placement
 
-- Target: reviewed Windows 11 worker with WSL2 Ubuntu 24.04.
+- Target: reviewed Windows 11 worker with the dedicated `MissionCore-Sim` WSL2
+  Ubuntu 24.04 distribution physically installed on D.
 - All mutable bytes must be under `D:\NDC_MISSIONCORE`.
-- Do not place a WSL distribution, Docker VHD, source, build, cache, log, bag,
-  ULog, Gazebo artifact or dataset on C.
+- Do not place a WSL distribution, source, build, cache, log, bag, ULog,
+  Gazebo artifact or dataset on C.
+- Do not use or reconfigure the existing Docker Desktop runtime in S0. Its
+  backend is C-backed and serves unrelated co-tenants.
 - Do not expose S0 services beyond loopback.
+- Run providers inside the ephemeral loopback-only network namespace created
+  by `simulation/s0/worker-smoke.sh`. Micro XRCE-DDS Agent binds UDP wildcard
+  inside that namespace because v2.4.3 has no address-selection option.
 - Do not download full external datasets.
 - Keep `actuator_authority=false`.
 
@@ -24,13 +30,34 @@ uv run missioncore-sim s0 doctor --json
 
 Before target evidence exists, the expected verdict is `INCOMPLETE`.
 
+## Stock-rover smoke
+
+Copy the reviewed script to the D-only worker source root, then invoke it as
+root inside `MissionCore-Sim` with a unique immutable run ID:
+
+```bash
+simulation/s0/worker-smoke.sh <UTC-run-id>-stock-rover
+```
+
+The script:
+
+- creates a new network namespace containing only loopback;
+- starts Micro XRCE-DDS Agent, PX4 SITL and the stock Gazebo Ackermann world;
+- verifies the DDS session, `/fmu/out/vehicle_status_v1` and Gazebo `/clock`;
+- records listeners and a non-acceptance resource point sample;
+- sends the PX4 `shutdown` command and rejects owned-process residue.
+
+Do not reuse a run ID. A failed run remains diagnostic evidence; rerun with a
+new ID after correction.
+
 ## Required target inventory
 
 Record a redacted artifact for each of:
 
 1. Windows release, WSL version, Ubuntu version and architecture.
 2. GPU/driver, CPU, RAM and VRAM.
-3. Physical locations of the WSL distribution and Docker data/VHD.
+3. Physical location of the dedicated WSL distribution and confirmation that
+   the Docker-free S0 profile is active.
 4. D total/free bytes before installation.
 5. Existing Docker images, WSL distributions and simulation-related processes.
 6. Existing listeners and the candidate port registry.