devagent-physical-engine 1.9.0__tar.gz → 1.9.1__tar.gz

{devagent_physical_engine-1.9.0/src/devagent_physical_engine.egg-info → devagent_physical_engine-1.9.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: devagent-physical-engine
-Version: 1.9.0
+Version: 1.9.1
 Summary: Verification-first robotics commissioning runtime with declarative customer report tests, case-centric evidence reuse, qualified UR5e Gazebo/MoveIt execution, exact-Twin lineage, selective verification, regression, replay, and FAT reporting.
 Author: Tom Ha
 License: All Rights Reserved
@@ -110,7 +110,7 @@ regression + FAT report + Evidence Viewer
 Python 3.11+ is required.
 
 ```bash
-python -m pip install --upgrade devagent-physical-engine==1.9.0
+python -m pip install --upgrade devagent-physical-engine==1.9.1
 ```
 
 Verify the installed package:
@@ -122,7 +122,7 @@ python -c "import devagent_physical_engine as d; print(d.__version__)"
 Expected:
 
 ```text
-1.9.0
+1.9.1
 ```
 
 Run the normal customer workflow:
@@ -552,6 +552,32 @@ Hosted regression includes an arbitrary customer report test using `FAT-CUSTOM-9
 
 Full release notes: https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_0.md
 
+## v1.9.1 repeatable fresh physical simulation
+
+v1.9.1 makes evidence reuse and operator-requested fresh execution explicit without changing physical authority boundaries.
+
+```text
+devagent> simulate TEST-PERF-001
+  → reuse exact qualified case evidence when available
+  → execute only when the mapped case evidence is missing
+
+devagent> simulate --fresh TEST-PERF-001
+  → force a new execution of the selected unique physical case
+  → append a new immutable measurement artifact
+
+devagent> simulate --fresh all
+  → force every unique mapped physical case once
+  → still deduplicate shared case IDs across report criteria
+```
+
+`rerun TEST-X`, `rerun failed`, `rerun blocked`, `rerun affected`, and `rerun all` use the same fresh-execution contract for physical selectors. Deterministic selectors remain evidence-only and never command Gazebo.
+
+Interactive verification now owns its Gazebo/MoveIt lifecycle. After an executed verification case captures its evidence, that verification-owned stack is closed so the next fresh command starts from a clean runtime instead of competing with a stale simulator in the same ROS domain. Persistent visualization is a separate intent and remains available through `devagent robot visualize`.
+
+The lifecycle change does not relax MoveIt preexecution, canonical start-state compatibility, FollowJointTrajectory result handling, controller-feedback tracking, Twin lineage, site qualification, or real-robot execution locks.
+
+Full release notes: https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_1.md
+
 ## Qualified visual-execution scope
 
 Measured report-test execution remains intentionally narrow:
@@ -575,11 +601,15 @@ Use `--interactive`, or `--open` from a real terminal, to keep a bounded prompt:
 status
 tests
 test SELECTOR
-simulate [SELECTOR]
-rerun [SELECTOR]
+simulate SELECTOR
+simulate all
+simulate --fresh SELECTOR
+simulate --fresh all
+rerun SELECTOR
 rerun failed
 rerun blocked
 rerun affected
+rerun all
 replay
 report
 log
@@ -604,6 +634,8 @@ all
 
 `simulate all` creates a report-wide capability execution plan, evaluates deterministic rows directly, deduplicates physical cases, reuses exact qualified evidence first, and executes only missing unique qualified cases.
 
+`simulate --fresh TEST-X` and `rerun TEST-X` explicitly request a new physical execution; fresh report-wide execution still deduplicates physical case IDs.
+
 `replay` reconstructs persisted evidence identity; it never commands Gazebo, MoveIt, a robot, PLC, or safety PLC.
 
 ## Customer input
@@ -754,11 +786,11 @@ devagent-physical-ai optional provider-backed engineering front end
 
 # Project status
 
-**v1.9.0 — Production/Stable software workflow; dynamic customer report verification with target-proven qualified UR5e physical baseline**
+**v1.9.1 — Production/Stable software workflow; repeatable fresh physical simulation with owned verification lifecycle**
 
-v1.9.0 adds data-driven report verification on top of the target-proven v1.8.9 physical runtime. Customer test IDs are opaque, report behavior is routed from declarative capability fields, arbitrary finite typed physical metrics can flow through evidence when a qualified adapter measures them, and report-wide simulation deduplicates physical cases so one qualified case measurement can satisfy multiple compatible KPIs without repeated robot motion.
+v1.9.1 preserves v1.9.0's declarative customer report verification and case-centric evidence reuse, while adding explicit fresh physical execution and deterministic simulator lifecycle ownership. Normal `simulate` remains evidence-first; `simulate --fresh ...` and physical `rerun ...` force new execution, and report-wide fresh runs still execute each unique mapped case only once.
 
-The physical qualification claim itself remains grounded in the v1.8.9 target acceptance: the exact candidate completed the bounded UR5e nominal report lifecycle on the intended workstation with exact Twin readback, Pilz preflight, verified GRASP, scoped support contact, Gazebo carry, transfer/place, physical release/settle, explicit MoveIt detach/world reassert, retreat, and final verification.
+The physical qualification claim remains grounded in the target-proven bounded UR5e nominal report lifecycle. v1.9.1 does not expand that claim to arbitrary customer cells and does not grant site qualification, functional-safety certification, or autonomous real-robot execution authority.
 
 Production-oriented software capabilities include:
 
@@ -774,6 +806,8 @@ requirement-driven nominal / boundary / fault case generation
 immutable project/Twin lineage
 exact Requirement → Mapping → Verification Plan → Case provenance
 case-centric qualified evidence reuse
+explicit fresh physical execution
+verification-owned simulator lifecycle cleanup
 report-wide physical-case deduplication
 mixed deterministic + physical `simulate all` orchestration
 truthful NOT_TESTED for unavailable/unsupported evidence
@@ -809,6 +843,9 @@ Production/Stable describes the bounded software workflow. The target-proven UR5
 
 ## Documentation
 
+v1.9.1 release notes:
+https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_1.md
+
 v1.9.0 release notes:
 https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_0.md
 

{devagent_physical_engine-1.9.0 → devagent_physical_engine-1.9.1}/README.md

@@ -68,7 +68,7 @@ regression + FAT report + Evidence Viewer
 Python 3.11+ is required.
 
 ```bash
-python -m pip install --upgrade devagent-physical-engine==1.9.0
+python -m pip install --upgrade devagent-physical-engine==1.9.1
 ```
 
 Verify the installed package:
@@ -80,7 +80,7 @@ python -c "import devagent_physical_engine as d; print(d.__version__)"
 Expected:
 
 ```text
-1.9.0
+1.9.1
 ```
 
 Run the normal customer workflow:
@@ -510,6 +510,32 @@ Hosted regression includes an arbitrary customer report test using `FAT-CUSTOM-9
 
 Full release notes: https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_0.md
 
+## v1.9.1 repeatable fresh physical simulation
+
+v1.9.1 makes evidence reuse and operator-requested fresh execution explicit without changing physical authority boundaries.
+
+```text
+devagent> simulate TEST-PERF-001
+  → reuse exact qualified case evidence when available
+  → execute only when the mapped case evidence is missing
+
+devagent> simulate --fresh TEST-PERF-001
+  → force a new execution of the selected unique physical case
+  → append a new immutable measurement artifact
+
+devagent> simulate --fresh all
+  → force every unique mapped physical case once
+  → still deduplicate shared case IDs across report criteria
+```
+
+`rerun TEST-X`, `rerun failed`, `rerun blocked`, `rerun affected`, and `rerun all` use the same fresh-execution contract for physical selectors. Deterministic selectors remain evidence-only and never command Gazebo.
+
+Interactive verification now owns its Gazebo/MoveIt lifecycle. After an executed verification case captures its evidence, that verification-owned stack is closed so the next fresh command starts from a clean runtime instead of competing with a stale simulator in the same ROS domain. Persistent visualization is a separate intent and remains available through `devagent robot visualize`.
+
+The lifecycle change does not relax MoveIt preexecution, canonical start-state compatibility, FollowJointTrajectory result handling, controller-feedback tracking, Twin lineage, site qualification, or real-robot execution locks.
+
+Full release notes: https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_1.md
+
 ## Qualified visual-execution scope
 
 Measured report-test execution remains intentionally narrow:
@@ -533,11 +559,15 @@ Use `--interactive`, or `--open` from a real terminal, to keep a bounded prompt:
 status
 tests
 test SELECTOR
-simulate [SELECTOR]
-rerun [SELECTOR]
+simulate SELECTOR
+simulate all
+simulate --fresh SELECTOR
+simulate --fresh all
+rerun SELECTOR
 rerun failed
 rerun blocked
 rerun affected
+rerun all
 replay
 report
 log
@@ -562,6 +592,8 @@ all
 
 `simulate all` creates a report-wide capability execution plan, evaluates deterministic rows directly, deduplicates physical cases, reuses exact qualified evidence first, and executes only missing unique qualified cases.
 
+`simulate --fresh TEST-X` and `rerun TEST-X` explicitly request a new physical execution; fresh report-wide execution still deduplicates physical case IDs.
+
 `replay` reconstructs persisted evidence identity; it never commands Gazebo, MoveIt, a robot, PLC, or safety PLC.
 
 ## Customer input
@@ -712,11 +744,11 @@ devagent-physical-ai optional provider-backed engineering front end
 
 # Project status
 
-**v1.9.0 — Production/Stable software workflow; dynamic customer report verification with target-proven qualified UR5e physical baseline**
+**v1.9.1 — Production/Stable software workflow; repeatable fresh physical simulation with owned verification lifecycle**
 
-v1.9.0 adds data-driven report verification on top of the target-proven v1.8.9 physical runtime. Customer test IDs are opaque, report behavior is routed from declarative capability fields, arbitrary finite typed physical metrics can flow through evidence when a qualified adapter measures them, and report-wide simulation deduplicates physical cases so one qualified case measurement can satisfy multiple compatible KPIs without repeated robot motion.
+v1.9.1 preserves v1.9.0's declarative customer report verification and case-centric evidence reuse, while adding explicit fresh physical execution and deterministic simulator lifecycle ownership. Normal `simulate` remains evidence-first; `simulate --fresh ...` and physical `rerun ...` force new execution, and report-wide fresh runs still execute each unique mapped case only once.
 
-The physical qualification claim itself remains grounded in the v1.8.9 target acceptance: the exact candidate completed the bounded UR5e nominal report lifecycle on the intended workstation with exact Twin readback, Pilz preflight, verified GRASP, scoped support contact, Gazebo carry, transfer/place, physical release/settle, explicit MoveIt detach/world reassert, retreat, and final verification.
+The physical qualification claim remains grounded in the target-proven bounded UR5e nominal report lifecycle. v1.9.1 does not expand that claim to arbitrary customer cells and does not grant site qualification, functional-safety certification, or autonomous real-robot execution authority.
 
 Production-oriented software capabilities include:
 
@@ -732,6 +764,8 @@ requirement-driven nominal / boundary / fault case generation
 immutable project/Twin lineage
 exact Requirement → Mapping → Verification Plan → Case provenance
 case-centric qualified evidence reuse
+explicit fresh physical execution
+verification-owned simulator lifecycle cleanup
 report-wide physical-case deduplication
 mixed deterministic + physical `simulate all` orchestration
 truthful NOT_TESTED for unavailable/unsupported evidence
@@ -767,6 +801,9 @@ Production/Stable describes the bounded software workflow. The target-proven UR5
 
 ## Documentation
 
+v1.9.1 release notes:
+https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_1.md
+
 v1.9.0 release notes:
 https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_0.md
 
@@ -802,4 +839,4 @@ Original creator: Tom Ha
 Original project: https://github.com/tomha85/devagent-physical-engine  
 All rights reserved.
 
-See repository `LICENSE`, `NOTICE`, and `COPYRIGHT` for complete ownership and usage terms.
+See repository `LICENSE`, `NOTICE`, and `COPYRIGHT` for complete ownership and usage terms.

{devagent_physical_engine-1.9.0 → devagent_physical_engine-1.9.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "devagent-physical-engine"
-version = "1.9.0"
+version = "1.9.1"
 description = "Verification-first robotics commissioning runtime with declarative customer report tests, case-centric evidence reuse, qualified UR5e Gazebo/MoveIt execution, exact-Twin lineage, selective verification, regression, replay, and FAT reporting."
 readme = { file = "README.md", content-type = "text/markdown" }
 requires-python = ">=3.11"
@@ -48,7 +48,7 @@ Issues = "https://github.com/tomha85/devagent-physical-engine/issues"
 Sponsor = "https://github.com/sponsors/tomha85"
 
 [project.scripts]
-devagent = "devagent_physical_engine.user_cli_v18:main"
+devagent = "devagent_physical_engine.user_cli_v191:main"
 devagent-physical = "devagent_physical_engine.cli:main"
 devagent-physical-ai = "devagent_physical_engine.ai_cli:main"
 devagent-commercial = "devagent_physical_engine.commercial_cli:main"

{devagent_physical_engine-1.9.0 → devagent_physical_engine-1.9.1}/src/devagent_physical_engine/__init__.py

@@ -219,4 +219,4 @@ install_industrial_motion_v170()
 install_controller_runtime_hotfix_v181()
 install_one_command_runtime()
 
-__version__ = "1.9.0"
+__version__ = "1.9.1"

devagent_physical_engine-1.9.1/src/devagent_physical_engine/user_cli_v191.py

@@ -0,0 +1,194 @@
+from __future__ import annotations
+
+import sys
+
+from . import user_cli_v16 as _v16
+from . import user_cli_v18 as _v18
+
+
+class InteractiveSessionV191(_v16.InteractiveSessionV16):
+    """Interactive console with explicit reuse-vs-fresh physical execution semantics.
+
+    ``simulate SELECTOR`` keeps the v1.9 evidence-first behavior: an exact qualified
+    case measurement is reused when one already exists. ``simulate --fresh SELECTOR``
+    and ``rerun SELECTOR`` express a different operator intent: execute each selected
+    unique physical case again and append new immutable evidence. Deterministic rows
+    remain evidence-only and never launch a simulator.
+
+    Interactive verification owns the simulator lifecycle for every executed case. A
+    successful case is therefore closed after evidence capture instead of leaving an
+    unmanaged Gazebo/MoveIt stack behind for the next interactive command. Persistent
+    visualization belongs to ``devagent robot visualize`` rather than verification.
+    """
+
+    def _executor(self, repository):
+        # Verification must be repeatable. Keeping a successful simulator open makes the
+        # next fresh command launch a second Gazebo/MoveIt stack into the same ROS domain,
+        # which can produce stale model read-back or startup-timeout failures. The case
+        # runner remains visible while it executes, but it is stopped after evidence is
+        # captured so the next command starts from a clean owned lifecycle.
+        return _v16.ReportTestSelectiveExecutor(
+            repository,
+            source_root=self.source,
+            keep_open=False,
+            progress=lambda stage, message: self.tracker.emit(
+                stage, "RUNNING", message
+            ),
+        )
+
+    def _fresh_simulation(self, selector: str, *, action: str) -> None:
+        self.tracker.emit(
+            "SELECTIVE",
+            "RUNNING",
+            f"fresh visual {action} selector={selector}",
+        )
+        repository = self._repository()
+        try:
+            artifact = self._service(repository).run_simulation(
+                project_id=self.result.project_id,
+                selector=selector,
+                parent_campaign_artifact_id=self.result.campaign_artifact_id,
+                executor=self._executor(repository),
+                force_execution=True,
+            )
+        except (_v16.CommercialError, _v16.ProjectError, OSError) as exc:
+            self.tracker.emit("SELECTIVE", "BLOCKED", str(exc))
+            print(f"Fresh selective simulation blocked: {exc}")
+            print("No unqualified simulator or robot command was substituted.")
+            return
+
+        state = str(artifact.payload.get("state") or "complete").upper()
+        reason = str(artifact.payload.get("reason") or "")
+        self.tracker.emit(
+            "SELECTIVE",
+            state,
+            f"artifact={artifact.artifact_id} selector={selector} fresh=true"
+            + (f" reason={reason}" if reason else ""),
+        )
+        self._refresh_viewer(repository)
+        self._print_selective(artifact.payload)
+        if reason == "report_capability_execution_plan_v190":
+            self._print_batch_plan(artifact.payload)
+
+        if artifact.payload.get("state") == "blocked":
+            print("No unqualified simulator or robot command was executed.")
+            return
+        if reason == "simulation_not_required_deterministic_evidence":
+            print(
+                "Fresh execution is not applicable to this deterministic test; "
+                "immutable Twin evidence was evaluated directly."
+            )
+            return
+
+        contracts = artifact.payload.get("contracts")
+        physical_selected = (
+            any(
+                isinstance(item, dict) and item.get("mode") == "physical"
+                for item in contracts
+            )
+            if isinstance(contracts, list)
+            else False
+        )
+        if physical_selected and not artifact.payload.get("execution_performed"):
+            # A fresh physical request must never silently degrade into evidence reuse.
+            self.tracker.emit(
+                "SELECTIVE",
+                "ERROR",
+                f"fresh execution contract not satisfied selector={selector}",
+            )
+            print(
+                "Fresh physical execution was requested but no physical case was "
+                "executed. Treat this result as invalid and inspect RUN.log."
+            )
+            return
+
+        if artifact.payload.get("execution_performed"):
+            print(
+                "Fresh physical execution completed; prior qualified measurements were "
+                "not reused for the requested case execution."
+            )
+            print(
+                "The new measurement is appended as immutable evidence; previous "
+                "measurements remain available for audit/replay."
+            )
+            print(
+                "The verification-owned Gazebo/MoveIt runtime was closed after evidence "
+                "capture so the next execution starts cleanly."
+            )
+
+    def _simulate(self, extra: list[str]) -> None:
+        fresh_count = sum(item == "--fresh" for item in extra)
+        if not fresh_count:
+            super()._simulate(extra)
+            return
+        selectors = [item for item in extra if item != "--fresh"]
+        if fresh_count != 1 or len(selectors) != 1 or len(extra) != 2:
+            print("Usage: simulate [--fresh] SELECTOR")
+            return
+        self._fresh_simulation(selectors[0], action="simulate")
+
+    def _rerun(self, extra: list[str]) -> None:
+        if len(extra) > 1:
+            print("Usage: rerun [SELECTOR]")
+            return
+        if extra:
+            self._fresh_simulation(extra[0], action="rerun")
+            return
+        self._verify_again(
+            simulate=self.result.simulation_requested,
+            action="RERUN",
+        )
+
+    @staticmethod
+    def _help() -> None:
+        print(
+            """Commands:
+  status                        show the latest bounded full-verification summary
+  tests                         show the live report matrix and evidence authority
+  test SELECTOR                 evaluate selected criteria from existing evidence only
+  simulate SELECTOR             reuse exact qualified case evidence, else execute missing case(s)
+  simulate all                  dedupe/reuse report cases and execute only missing case(s)
+  simulate --fresh SELECTOR     force fresh execution of selected unique physical case(s)
+  simulate --fresh all          force every unique mapped physical case once
+  rerun SELECTOR                alias for fresh selective execution for physical selectors
+  rerun failed                  freshly execute physical cases behind failed criteria
+  rerun blocked                 freshly execute physical cases behind NOT_TESTED/blocked criteria
+  rerun affected                freshly execute physical cases selected by change impact
+  rerun all                     freshly execute every unique mapped physical case once
+  rerun                         repeat the prior full-verification mode
+  replay                        inspect/reconstruct stored evidence; never commands a robot
+  report                        open the current Evidence Viewer in a browser tab
+  log [N]                       show the last N RUN.log lines (default 40, maximum 500)
+  help                          show this command list
+  quit                          leave the interactive session
+
+A report TEST-ID is a verification criterion, not a robot-run identity. Multiple tests
+may consume the same physical case measurement. Normal `simulate` therefore reuses exact
+qualified evidence by design. Use `--fresh` or `rerun SELECTOR` when the operator
+explicitly wants a new physical execution. Fresh report-wide execution still deduplicates
+case IDs so one physical scenario is never repeated merely because several criteria
+consume its evidence. Deterministic tests remain evidence-only and never launch Gazebo.
+Verification-owned Gazebo/MoveIt sessions are closed after each executed case; use
+`devagent robot visualize` when a persistent visual session is desired.
+"""
+        )
+
+
+def _run_v191(argv: list[str]) -> int:
+    original_session = _v16.InteractiveSessionV16
+    _v16.InteractiveSessionV16 = InteractiveSessionV191
+    try:
+        return _v16.main(argv)
+    finally:
+        _v16.InteractiveSessionV16 = original_session
+
+
+def main(argv: list[str] | None = None) -> int:
+    effective = list(sys.argv[1:] if argv is None else argv)
+    if effective and effective[0] in {"robots", "robot"}:
+        return _v18.main(effective)
+    return _run_v191(effective)
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

{devagent_physical_engine-1.9.0 → devagent_physical_engine-1.9.1/src/devagent_physical_engine.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: devagent-physical-engine
-Version: 1.9.0
+Version: 1.9.1
 Summary: Verification-first robotics commissioning runtime with declarative customer report tests, case-centric evidence reuse, qualified UR5e Gazebo/MoveIt execution, exact-Twin lineage, selective verification, regression, replay, and FAT reporting.
 Author: Tom Ha
 License: All Rights Reserved
@@ -110,7 +110,7 @@ regression + FAT report + Evidence Viewer
 Python 3.11+ is required.
 
 ```bash
-python -m pip install --upgrade devagent-physical-engine==1.9.0
+python -m pip install --upgrade devagent-physical-engine==1.9.1
 ```
 
 Verify the installed package:
@@ -122,7 +122,7 @@ python -c "import devagent_physical_engine as d; print(d.__version__)"
 Expected:
 
 ```text
-1.9.0
+1.9.1
 ```
 
 Run the normal customer workflow:
@@ -552,6 +552,32 @@ Hosted regression includes an arbitrary customer report test using `FAT-CUSTOM-9
 
 Full release notes: https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_0.md
 
+## v1.9.1 repeatable fresh physical simulation
+
+v1.9.1 makes evidence reuse and operator-requested fresh execution explicit without changing physical authority boundaries.
+
+```text
+devagent> simulate TEST-PERF-001
+  → reuse exact qualified case evidence when available
+  → execute only when the mapped case evidence is missing
+
+devagent> simulate --fresh TEST-PERF-001
+  → force a new execution of the selected unique physical case
+  → append a new immutable measurement artifact
+
+devagent> simulate --fresh all
+  → force every unique mapped physical case once
+  → still deduplicate shared case IDs across report criteria
+```
+
+`rerun TEST-X`, `rerun failed`, `rerun blocked`, `rerun affected`, and `rerun all` use the same fresh-execution contract for physical selectors. Deterministic selectors remain evidence-only and never command Gazebo.
+
+Interactive verification now owns its Gazebo/MoveIt lifecycle. After an executed verification case captures its evidence, that verification-owned stack is closed so the next fresh command starts from a clean runtime instead of competing with a stale simulator in the same ROS domain. Persistent visualization is a separate intent and remains available through `devagent robot visualize`.
+
+The lifecycle change does not relax MoveIt preexecution, canonical start-state compatibility, FollowJointTrajectory result handling, controller-feedback tracking, Twin lineage, site qualification, or real-robot execution locks.
+
+Full release notes: https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_1.md
+
 ## Qualified visual-execution scope
 
 Measured report-test execution remains intentionally narrow:
@@ -575,11 +601,15 @@ Use `--interactive`, or `--open` from a real terminal, to keep a bounded prompt:
 status
 tests
 test SELECTOR
-simulate [SELECTOR]
-rerun [SELECTOR]
+simulate SELECTOR
+simulate all
+simulate --fresh SELECTOR
+simulate --fresh all
+rerun SELECTOR
 rerun failed
 rerun blocked
 rerun affected
+rerun all
 replay
 report
 log
@@ -604,6 +634,8 @@ all
 
 `simulate all` creates a report-wide capability execution plan, evaluates deterministic rows directly, deduplicates physical cases, reuses exact qualified evidence first, and executes only missing unique qualified cases.
 
+`simulate --fresh TEST-X` and `rerun TEST-X` explicitly request a new physical execution; fresh report-wide execution still deduplicates physical case IDs.
+
 `replay` reconstructs persisted evidence identity; it never commands Gazebo, MoveIt, a robot, PLC, or safety PLC.
 
 ## Customer input
@@ -754,11 +786,11 @@ devagent-physical-ai optional provider-backed engineering front end
 
 # Project status
 
-**v1.9.0 — Production/Stable software workflow; dynamic customer report verification with target-proven qualified UR5e physical baseline**
+**v1.9.1 — Production/Stable software workflow; repeatable fresh physical simulation with owned verification lifecycle**
 
-v1.9.0 adds data-driven report verification on top of the target-proven v1.8.9 physical runtime. Customer test IDs are opaque, report behavior is routed from declarative capability fields, arbitrary finite typed physical metrics can flow through evidence when a qualified adapter measures them, and report-wide simulation deduplicates physical cases so one qualified case measurement can satisfy multiple compatible KPIs without repeated robot motion.
+v1.9.1 preserves v1.9.0's declarative customer report verification and case-centric evidence reuse, while adding explicit fresh physical execution and deterministic simulator lifecycle ownership. Normal `simulate` remains evidence-first; `simulate --fresh ...` and physical `rerun ...` force new execution, and report-wide fresh runs still execute each unique mapped case only once.
 
-The physical qualification claim itself remains grounded in the v1.8.9 target acceptance: the exact candidate completed the bounded UR5e nominal report lifecycle on the intended workstation with exact Twin readback, Pilz preflight, verified GRASP, scoped support contact, Gazebo carry, transfer/place, physical release/settle, explicit MoveIt detach/world reassert, retreat, and final verification.
+The physical qualification claim remains grounded in the target-proven bounded UR5e nominal report lifecycle. v1.9.1 does not expand that claim to arbitrary customer cells and does not grant site qualification, functional-safety certification, or autonomous real-robot execution authority.
 
 Production-oriented software capabilities include:
 
@@ -774,6 +806,8 @@ requirement-driven nominal / boundary / fault case generation
 immutable project/Twin lineage
 exact Requirement → Mapping → Verification Plan → Case provenance
 case-centric qualified evidence reuse
+explicit fresh physical execution
+verification-owned simulator lifecycle cleanup
 report-wide physical-case deduplication
 mixed deterministic + physical `simulate all` orchestration
 truthful NOT_TESTED for unavailable/unsupported evidence
@@ -809,6 +843,9 @@ Production/Stable describes the bounded software workflow. The target-proven UR5
 
 ## Documentation
 
+v1.9.1 release notes:
+https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_1.md
+
 v1.9.0 release notes:
 https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V1_9_0.md
 

{devagent_physical_engine-1.9.0 → devagent_physical_engine-1.9.1}/src/devagent_physical_engine.egg-info/SOURCES.txt

@@ -71,6 +71,7 @@ src/devagent_physical_engine/user_cli.py
 src/devagent_physical_engine/user_cli_v15.py
 src/devagent_physical_engine/user_cli_v16.py
 src/devagent_physical_engine/user_cli_v18.py
+src/devagent_physical_engine/user_cli_v191.py
 src/devagent_physical_engine/verification.py
 src/devagent_physical_engine/verification_capabilities_v190.py
 src/devagent_physical_engine/workpiece_lifecycle.py
@@ -202,6 +203,8 @@ tests/test_evidence_trust_transport_v12.py
 tests/test_evidence_trust_v12.py
 tests/test_evidence_viewer_palette_v130.py
 tests/test_explicit_detach_world_reassert_v189.py
+tests/test_fresh_simulation_cli_v191.py
+tests/test_fresh_simulation_service_v191.py
 tests/test_gazebo_cli_v0122.py
 tests/test_golden_abb_runner_v171.py
 tests/test_golden_abb_showcase.py

{devagent_physical_engine-1.9.0 → devagent_physical_engine-1.9.1}/src/devagent_physical_engine.egg-info/entry_points.txt

@@ -1,5 +1,5 @@
 [console_scripts]
-devagent = devagent_physical_engine.user_cli_v18:main
+devagent = devagent_physical_engine.user_cli_v191:main
 devagent-commercial = devagent_physical_engine.commercial_cli:main
 devagent-physical = devagent_physical_engine.cli:main
 devagent-physical-ai = devagent_physical_engine.ai_cli:main