RMANOV · RMANOV · Jun 15, 2026 · Jun 15, 2026
@@ -0,0 +1,197 @@
+# STRIX Reviewer Demo — `gps_denied_recon` (Software Replay)
+
+> Simulation-only. Zero hardware flights. No RF/sensor fidelity. No field validation. Safety claims are limited to classical-CBF + ROE + traces + simulator; no trained-neural safety guarantee is claimed.
+
+This is a **static, reviewer-facing** walkthrough of one deterministic software-only
+scenario replay. It shows exactly how to reproduce a single worked scenario and what
+the harness emits, so a reviewer can re-run it and obtain byte-identical output.
+
+- **Scope:** documentation-only. No source, manifest, or test was modified to produce this doc.
+- **Evidence kind:** `software_replay`, `fidelity = deterministic_kinematic_public_replay`.
+- **As-of commit:** `3a5d46de4ed69f4321c058ae55d84047e2ccad49` (`main`, working tree clean).
+- **Claim posture:** every figure below is a *prior measured software-replay result to be
+  re-run on the exact submission commit*, never a final live fact, and never inferred
+  hardware/RF/field readiness. See [`../CAPABILITY_BOUNDARY.md`](../CAPABILITY_BOUNDARY.md).
+
+The replay harness is a lightweight **kinematic** scenario player. It is **not** a
+field-physics, RF, or hardware-fidelity simulator. Its purpose is repeatable,
+inspectable public scenario behavior before any integration testing.
+
+---
+
+## 1. The worked scenario
+
+`gps_denied_recon` — four aerial agents fly a reconnaissance pattern and transition to
+GPS-denied (degraded) navigation when GPS is lost at `t = 30 s`. The scenario's seed is
+fixed at **42001** inside the scenario file itself
+(`sim/scenarios/gps_denied_recon.yaml`), so the replay is deterministic by construction.
+
+```
+Scenario:  gps_denied_recon
+Seed:      42001            (defined in the scenario YAML; the harness reads it)
+Duration:  600 s           Tick: 10 s            Frames: 61            Agents: 4
+```
+
+---
+
+## 2. Exact replay command
+
+The harness takes the scenario file as input and reads `seed: 42001` from it (there is no
+`--seed` flag — the seed is a scenario property). Run from the repository root:
+
+```bash
+python3 scripts/strix_sim_replay.py \
+  --scenario sim/scenarios/gps_denied_recon.yaml \
+  --output /tmp/a1_run.json \
+  --no-html
+```
+
+`--no-html` writes only the JSON evidence. The process exits `0` when the scenario's
+pass-envelope is satisfied.
+
+---
+
+## 3. Determinism proof (double run, byte-identical)
+
+The command was run **twice** in sequence into two separate files, then compared byte-for-byte:
+
+```bash
+python3 scripts/strix_sim_replay.py --scenario sim/scenarios/gps_denied_recon.yaml --output /tmp/a1_run1.json --no-html
+python3 scripts/strix_sim_replay.py --scenario sim/scenarios/gps_denied_recon.yaml --output /tmp/a1_run2.json --no-html
+cmp -s /tmp/a1_run1.json /tmp/a1_run2.json && echo "byte-identical"
+```
+
+Within a single checkout, the two runs are **byte-for-byte identical to each other** —
+`cmp -s` reports no difference.
+
+> The whole-file `sha256sum` is **intentionally not pinned** in this doc. The replay
+> JSON embeds a `repo` block (`commit`, `branch`, `working_tree_clean`) that reflects
+> *your* git context, not the simulation, so the same seeded run yields a *different*
+> whole-file hash per checkout — any concrete value here would be a checkout-specific
+> reviewer trap.
+>
+> What **does** reproduce deterministically on any checkout is the seed-driven
+> simulation content — the `scenario`, `metrics`, and `envelope` blocks (Section 4).
+> Verify reproducibility by comparing **those blocks**, not a whole-file SHA.
+
+---
+
+## 4. Captured output excerpt
+
+A captured excerpt of the harness output for this run is stored alongside this doc at
+[`replay_output.txt`](replay_output.txt). The load-bearing parts are reproduced below.
+
+### 4.1 Scenario header — `replay["scenario"]`
+
+```json
+{
+  "config_hash": "829cc17c46b98d72",
+  "duration_s": 600.0,
+  "id": "gps_denied_recon",
+  "name": "GPS-Denied Reconnaissance",
+  "path": "sim/scenarios/gps_denied_recon.yaml",
+  "seed": 42001,
+  "tick_s": 10.0
+}
+```
+
+### 4.2 Metrics — `replay["metrics"]`
+
+```json
+{
+  "active_agents": 4,
+  "area_coverage_pct": 100.0,
+  "formation_coherence": 0.78,
+  "frame_count": 61,
+  "mean_energy_remaining_pct": 66.111,
+  "min_constraint_clearance_m": 0.0,
+  "offline_agents": 0,
+  "position_error_rms_m": 4.5
+}
+```
+
+These are prior measured software-replay values for this seed/commit, to be re-run on the
+exact submission commit. They are not hardware, RF, or field measurements.
+
+### 4.3 Pass-envelope evaluation — `replay["envelope"]`
+
+```json
+{
+  "checks": [
+    { "metric": "area_coverage_pct",   "observed": 100.0, "min": 80,  "max": 100,  "status": "passed" },
+    { "metric": "position_error_rms_m", "observed": 4.5,   "min": 0.0, "max": 10.0, "status": "passed" },
+    { "metric": "formation_coherence",  "observed": 0.78,  "min": 0.6, "max": 1.0,  "status": "passed" }
+  ],
+  "status": "passed"
+}
+```
+
+### 4.4 Timeline excerpt — degraded-nav transition at GPS loss
+
+The per-agent `mode` field flips from `nominal` to `degraded_nav` exactly at the
+scheduled `t = 30 s` `gps_loss` event, and scenario events surface at their scheduled
+times. This is the documented degraded-mode / EW behavior under GPS loss.
+
+```
+  t_s  agent_1.mode      agent_1 (x, y, z)               events
+-----  ----------------  ------------------------------  ------------------------
+    0  nominal           (    0.00,   200.00,  -50.00)   -
+   10  nominal           (  -51.89,   218.94,  -50.00)   -
+   20  nominal           ( -100.98,   201.07,  -50.00)   -
+   30  degraded_nav      ( -144.39,   171.75,  -50.00)   30.0s:gps_loss
+   40  degraded_nav      ( -180.79,   134.60,  -50.00)   -
+  120  degraded_nav      (  -76.23,  -212.21,  -50.00)   120.0s:wind_gust
+  300  degraded_nav      ( -144.40,   172.08,  -50.00)   300.0s:sensor_degradation
+  600  degraded_nav      ( -222.22,    39.13,  -50.00)   -
+```
+
+This scenario defines no threat constraints, so `min_constraint_clearance_m` is `0.0`
+and no constraint-avoidance gating fires in this run. The shipped safety story rests on
+classical control-barrier-function gating plus ROE plus traces plus the simulator; this
+particular scenario exercises the GPS-denied degraded-navigation path, not CBF gating.
+
+---
+
+## 5. Diagrams (labels only)
+
+### 5.1 Replay data flow
+
+```mermaid
+flowchart LR
+    Y["gps_denied_recon.yaml<br/>seed=42001"] --> R["strix_sim_replay.py<br/>kinematic replay"]
+    R --> J["replay JSON<br/>scenario / metrics / envelope / frames"]
+    J --> E["pass-envelope check"]
+    J --> T["timeline / modes"]
+```
+
+### 5.2 Navigation-mode state (this scenario)
+
+```
+   t < 30s              t >= 30s (gps_loss)
+ +----------+          +---------------+
+ | nominal  | -------> | degraded_nav  |
+ +----------+          +---------------+
+```
+
+---
+
+## 6. Reproduce checklist
+
+1. Check out commit `3a5d46de4ed69f4321c058ae55d84047e2ccad49`, clean working tree.
+2. From the repo root, run the command in Section 2.
+3. Run it a second time to a different output path.
+4. `cmp -s` the two outputs — within your checkout they are byte-identical to each
+   other. (The whole-file `sha256sum` is checkout-specific and intentionally not
+   pinned — see Section 3.)
+5. Compare your `scenario` / `metrics` / `envelope` blocks to Section 4 — these are the
+   seed-deterministic, checkout-independent reproducibility check.
+
+---
+
+## 7. Boundary reminder
+
+- Software replay only. No hardware flights, no RF/sensor fidelity, no field validation.
+- Numbers are prior measured software-replay results, re-run on the submission commit.
+- No fielded deployment, no delivered ROS2/MAVLink hardware integration, no defence
+  validation/accreditation, no trained-neural safety guarantee is claimed.
+- Authoritative claim map: [`../CAPABILITY_BOUNDARY.md`](../CAPABILITY_BOUNDARY.md).
@@ -0,0 +1,81 @@
+STRIX software-replay -- captured deterministic output excerpt
+Scenario: gps_denied_recon   Seed: 42001
+Source: scripts/strix_sim_replay.py  (kind=software_replay, fidelity=deterministic_kinematic_public_replay)
+As-of commit: 3a5d46de4ed69f4321c058ae55d84047e2ccad49 (main, working tree clean)
+
+This file is an excerpt of the replay harness output: JSON blocks plus a derived/hand-formatted timeline and summary. Two consecutive
+runs in the same checkout produce a byte-identical file (cmp -s). The whole-file SHA256 is
+intentionally not pinned: the JSON embeds a git repo block, so the whole-file hash varies
+per checkout. The seed-driven scenario/metrics/envelope content is what reproduces
+deterministically across checkouts.
+
+--- scenario header (replay['scenario']) ---
+{
+  "config_hash": "829cc17c46b98d72",
+  "duration_s": 600.0,
+  "id": "gps_denied_recon",
+  "name": "GPS-Denied Reconnaissance",
+  "path": "sim/scenarios/gps_denied_recon.yaml",
+  "seed": 42001,
+  "tick_s": 10.0
+}
+
+--- metrics (replay['metrics']) ---
+{
+  "active_agents": 4,
+  "area_coverage_pct": 100.0,
+  "formation_coherence": 0.78,
+  "frame_count": 61,
+  "mean_energy_remaining_pct": 66.111,
+  "min_constraint_clearance_m": 0.0,
+  "offline_agents": 0,
+  "position_error_rms_m": 4.5
+}
+
+--- pass-envelope evaluation (replay['envelope']) ---
+{
+  "checks": [
+    {
+      "max": 100,
+      "metric": "area_coverage_pct",
+      "min": 80,
+      "observed": 100.0,
+      "status": "passed"
+    },
+    {
+      "max": 10.0,
+      "metric": "position_error_rms_m",
+      "min": 0.0,
+      "observed": 4.5,
+      "status": "passed"
+    },
+    {
+      "max": 1.0,
+      "metric": "formation_coherence",
+      "min": 0.6,
+      "observed": 0.78,
+      "status": "passed"
+    }
+  ],
+  "status": "passed"
+}
+
+--- timeline excerpt: agent_1 navigation mode at key timestamps ---
+(degraded_nav engages exactly at the t=30s scheduled gps_loss event)
+
+  t_s  agent_1.mode      agent_1 (x, y, z)               events
+-----  ----------------  ------------------------------  ------------------------
+    0  nominal           (    0.00,   200.00,  -50.00)   -
+   10  nominal           (  -51.89,   218.94,  -50.00)   -
+   20  nominal           ( -100.98,   201.07,  -50.00)   -
+   30  degraded_nav      ( -144.39,   171.75,  -50.00)   30.0s:gps_loss
+   40  degraded_nav      ( -180.79,   134.60,  -50.00)   -
+  120  degraded_nav      (  -76.23,  -212.21,  -50.00)   120.0s:wind_gust
+  300  degraded_nav      ( -144.40,   172.08,  -50.00)   300.0s:sensor_degradation
+  600  degraded_nav      ( -222.22,    39.13,  -50.00)   -
+
+Total frames: 61   Agents: 4   Constraints: 0
+
+Caveat: simulation-only deterministic kinematic replay. No RF/sensor fidelity,
+no hardware, no field validation. Safety claims are limited to classical-CBF + ROE + traces + simulator; no trained-neural safety guarantee is claimed.
+