diff --git a/crates/console/prometeu-firmware/src/firmware/firmware.rs b/crates/console/prometeu-firmware/src/firmware/firmware.rs
index 2c88ee1f..3d6bf562 100644
--- a/crates/console/prometeu-firmware/src/firmware/firmware.rs
+++ b/crates/console/prometeu-firmware/src/firmware/firmware.rs
@@ -177,6 +177,8 @@ mod tests {
     use prometeu_hal::cartridge::{AppMode, AssetsPayloadSource};
     use prometeu_hal::syscalls::caps;
     use prometeu_system::CrashReport;
+    use prometeu_system::process::ProcessState;
+    use prometeu_system::task::TaskState;
 
     fn halting_program() -> Vec<u8> {
         let code = assemble("HALT").expect("assemble");
@@ -289,7 +291,10 @@ mod tests {
         firmware.load_cartridge(trapping_game_cartridge());
         firmware.tick(&signals, &mut hardware);
 
-        assert!(matches!(firmware.state, FirmwareState::GameRunning(_)));
+        let task_id = match &firmware.state {
+            FirmwareState::GameRunning(step) => step.task_id,
+            other => panic!("expected GameRunning state, got {:?}", other),
+        };
 
         firmware.tick(&signals, &mut hardware);
 
@@ -302,6 +307,13 @@ mod tests {
             },
             other => panic!("expected AppCrashes state, got {:?}", other),
         }
+
+        let task = firmware.os.task_manager.get(task_id).expect("task should exist");
+        let process =
+            firmware.os.process_manager.get(task.process_id).expect("process should exist");
+
+        assert_eq!(task.state, TaskState::Crashed);
+        assert_eq!(process.state, ProcessState::Crashed);
     }
 
     #[test]
diff --git a/crates/console/prometeu-firmware/src/firmware/firmware_step_game_running.rs b/crates/console/prometeu-firmware/src/firmware/firmware_step_game_running.rs
index 63fdfda1..43d7ae63 100644
--- a/crates/console/prometeu-firmware/src/firmware/firmware_step_game_running.rs
+++ b/crates/console/prometeu-firmware/src/firmware/firmware_step_game_running.rs
@@ -19,19 +19,20 @@ impl GameRunningStep {
     }
 
     pub fn on_update(&mut self, ctx: &mut PrometeuContext) -> Option<FirmwareState> {
-        let Some(task) = ctx.os.task_manager.get(self.task_id) else {
+        let Some(task_state) = ctx.os.task_manager.get(self.task_id).map(|task| task.state) else {
             // TODO: is it panic?
             let report =
                 CrashReport::VmPanic { message: "task is gone, panic!".to_string(), pc: None };
             return Some(FirmwareState::AppCrashes(AppCrashesStep { report }));
         };
 
-        if task.state != TaskState::Foreground {
+        if task_state != TaskState::Foreground {
             // TODO: is it panic?
             let report = CrashReport::VmPanic {
                 message: "should be running as foreground, panic!".to_string(),
                 pc: None,
             };
+            let _ = ctx.os.crash_task(self.task_id, Some(&report));
             return Some(FirmwareState::AppCrashes(AppCrashesStep { report }));
         }
 
@@ -41,7 +42,12 @@ impl GameRunningStep {
             ctx.hw.gfx_mut().present();
         }
 
-        result.map(|report| FirmwareState::AppCrashes(AppCrashesStep { report }))
+        if let Some(report) = result {
+            let _ = ctx.os.crash_task(self.task_id, Some(&report));
+            return Some(FirmwareState::AppCrashes(AppCrashesStep { report }));
+        }
+
+        None
     }
 
     pub fn on_exit(&mut self, _ctx: &mut PrometeuContext) {}
diff --git a/crates/console/prometeu-firmware/src/firmware/firmware_step_system_running.rs b/crates/console/prometeu-firmware/src/firmware/firmware_step_system_running.rs
index d7dd7ab8..bcbccdf8 100644
--- a/crates/console/prometeu-firmware/src/firmware/firmware_step_system_running.rs
+++ b/crates/console/prometeu-firmware/src/firmware/firmware_step_system_running.rs
@@ -20,6 +20,7 @@ impl SystemRunningStep {
     pub fn on_update(&mut self, ctx: &mut PrometeuContext) -> Option<FirmwareState> {
         let outcome = ctx.hub.update_system_profile(ctx.os, ctx.vm, ctx.signals, ctx.hw);
         if let Some(report) = outcome.crash {
+            let _ = ctx.os.crash_task(self.task_id, Some(&report));
             return Some(FirmwareState::AppCrashes(AppCrashesStep { report }));
         }
 
diff --git a/crates/console/prometeu-system/src/lib.rs b/crates/console/prometeu-system/src/lib.rs
index 78a9ef7f..1773a865 100644
--- a/crates/console/prometeu-system/src/lib.rs
+++ b/crates/console/prometeu-system/src/lib.rs
@@ -4,7 +4,7 @@ mod programs;
 mod services;
 
 pub use crash_report::CrashReport;
-pub use os::SystemOS;
+pub use os::{LifecycleError, LifecycleOperation, SystemOS};
 pub use programs::PrometeuHub;
 pub use services::fs;
 pub use services::process;
diff --git a/crates/console/prometeu-system/src/os/lifecycle.rs b/crates/console/prometeu-system/src/os/lifecycle.rs
new file mode 100644
index 00000000..bd771d47
--- /dev/null
+++ b/crates/console/prometeu-system/src/os/lifecycle.rs
@@ -0,0 +1,18 @@
+use crate::process::ProcessId;
+use crate::task::{TaskId, TaskState};
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum LifecycleOperation {
+    SetForeground,
+    Suspend,
+    Resume,
+    Close,
+    Crash,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum LifecycleError {
+    TaskNotFound(TaskId),
+    ProcessNotFound(ProcessId),
+    InvalidTransition { task_id: TaskId, from: TaskState, operation: LifecycleOperation },
+}
diff --git a/crates/console/prometeu-system/src/os/mod.rs b/crates/console/prometeu-system/src/os/mod.rs
index 61995fd9..5a2e6d76 100644
--- a/crates/console/prometeu-system/src/os/mod.rs
+++ b/crates/console/prometeu-system/src/os/mod.rs
@@ -1,3 +1,5 @@
+mod lifecycle;
 mod system_os;
 
+pub use lifecycle::{LifecycleError, LifecycleOperation};
 pub use system_os::SystemOS;
diff --git a/crates/console/prometeu-system/src/os/system_os.rs b/crates/console/prometeu-system/src/os/system_os.rs
index 8f9fecee..ae536400 100644
--- a/crates/console/prometeu-system/src/os/system_os.rs
+++ b/crates/console/prometeu-system/src/os/system_os.rs
@@ -1,11 +1,12 @@
 use crate::CrashReport;
 use crate::VirtualMachineRuntime;
 use crate::fs::{FsBackend, FsState, VirtualFS};
+use crate::os::{LifecycleError, LifecycleOperation};
 use crate::process::ProcessManager;
 use crate::services::memcard::MemcardService;
 use crate::services::window_manager::WindowManager;
-use crate::task::TaskId;
 use crate::task::TaskManager;
+use crate::task::{TaskId, TaskState};
 use prometeu_hal::cartridge::Cartridge;
 use prometeu_hal::log::{LogLevel, LogService, LogSource};
 use prometeu_hal::telemetry::CertificationConfig;
@@ -102,7 +103,8 @@ impl SystemOS {
         let title = title.into();
         let process_id = self.process_manager.spawn_vm_game(app_id, title.clone());
         let task_id = self.task_manager.create_game_task(process_id, app_id, title);
-        self.task_manager.set_foreground(task_id);
+        self.set_foreground_task(task_id)
+            .expect("newly created game task should have an associated process");
 
         task_id
     }
@@ -111,7 +113,8 @@ impl SystemOS {
         let title = title.into();
         let process_id = self.process_manager.spawn_vm_shell(app_id, title.clone());
         let task_id = self.task_manager.create_shell_task(process_id, app_id, title);
-        self.task_manager.set_foreground(task_id);
+        self.set_foreground_task(task_id)
+            .expect("newly created shell task should have an associated process");
 
         task_id
     }
@@ -120,8 +123,202 @@ impl SystemOS {
         let title = title.into();
         let process_id = self.process_manager.spawn_native_shell(app_id, title.clone());
         let task_id = self.task_manager.create_shell_task(process_id, app_id, title);
-        self.task_manager.set_foreground(task_id);
+        self.set_foreground_task(task_id)
+            .expect("newly created native shell task should have an associated process");
 
         task_id
     }
+
+    pub fn set_foreground_task(&mut self, task_id: TaskId) -> Result<(), LifecycleError> {
+        let process_id = self.process_id_for_task(task_id)?;
+
+        self.task_manager.set_foreground(task_id);
+        self.process_manager.mark_running(process_id);
+
+        Ok(())
+    }
+
+    pub fn suspend_task(&mut self, task_id: TaskId) -> Result<(), LifecycleError> {
+        let process_id = self.process_id_for_task(task_id)?;
+
+        self.task_manager.mark_suspended(task_id);
+        self.process_manager.mark_suspended(process_id);
+
+        Ok(())
+    }
+
+    pub fn resume_task(&mut self, task_id: TaskId) -> Result<(), LifecycleError> {
+        let task = self.task_manager.get(task_id).ok_or(LifecycleError::TaskNotFound(task_id))?;
+
+        if task.state != TaskState::Suspended {
+            return Err(LifecycleError::InvalidTransition {
+                task_id,
+                from: task.state,
+                operation: LifecycleOperation::Resume,
+            });
+        }
+
+        let process_id = self.process_id_for_task(task_id)?;
+
+        self.task_manager.set_foreground(task_id);
+        self.process_manager.mark_running(process_id);
+
+        Ok(())
+    }
+
+    pub fn close_task(&mut self, task_id: TaskId) -> Result<(), LifecycleError> {
+        let process_id = self.process_id_for_task(task_id)?;
+
+        self.task_manager.close_task(task_id);
+        self.process_manager.mark_stopped(process_id);
+
+        Ok(())
+    }
+
+    pub fn crash_task(
+        &mut self,
+        task_id: TaskId,
+        _report: Option<&CrashReport>,
+    ) -> Result<(), LifecycleError> {
+        let process_id = self.process_id_for_task(task_id)?;
+
+        self.task_manager.mark_crashed(task_id);
+        self.process_manager.mark_crashed(process_id);
+
+        Ok(())
+    }
+
+    fn process_id_for_task(
+        &self,
+        task_id: TaskId,
+    ) -> Result<crate::process::ProcessId, LifecycleError> {
+        let task = self.task_manager.get(task_id).ok_or(LifecycleError::TaskNotFound(task_id))?;
+
+        if !self.process_manager.contains(task.process_id) {
+            return Err(LifecycleError::ProcessNotFound(task.process_id));
+        }
+
+        Ok(task.process_id)
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::process::ProcessState;
+
+    fn process_state_for_task(os: &SystemOS, task_id: TaskId) -> ProcessState {
+        let task = os.task_manager.get(task_id).expect("task should exist");
+        os.process_manager.get(task.process_id).expect("process should exist").state
+    }
+
+    #[test]
+    fn set_foreground_task_marks_task_and_process_running() {
+        let mut os = SystemOS::new(None);
+        let task_id = os.create_vm_game_task(42, "Sector Crawl");
+
+        os.suspend_task(task_id).expect("suspend should succeed");
+        os.set_foreground_task(task_id).expect("foreground should succeed");
+
+        assert_eq!(
+            os.task_manager.get(task_id).expect("task should exist").state,
+            TaskState::Foreground
+        );
+        assert_eq!(process_state_for_task(&os, task_id), ProcessState::Running);
+    }
+
+    #[test]
+    fn suspend_task_marks_task_and_process_suspended() {
+        let mut os = SystemOS::new(None);
+        let task_id = os.create_vm_game_task(42, "Sector Crawl");
+
+        os.suspend_task(task_id).expect("suspend should succeed");
+
+        assert_eq!(
+            os.task_manager.get(task_id).expect("task should exist").state,
+            TaskState::Suspended
+        );
+        assert_eq!(process_state_for_task(&os, task_id), ProcessState::Suspended);
+    }
+
+    #[test]
+    fn resume_task_marks_suspended_task_foreground_and_process_running() {
+        let mut os = SystemOS::new(None);
+        let task_id = os.create_vm_game_task(42, "Sector Crawl");
+
+        os.suspend_task(task_id).expect("suspend should succeed");
+        os.resume_task(task_id).expect("resume should succeed");
+
+        assert_eq!(
+            os.task_manager.get(task_id).expect("task should exist").state,
+            TaskState::Foreground
+        );
+        assert_eq!(process_state_for_task(&os, task_id), ProcessState::Running);
+    }
+
+    #[test]
+    fn close_task_marks_task_closed_and_process_stopped_without_removing_entities() {
+        let mut os = SystemOS::new(None);
+        let task_id = os.create_vm_game_task(42, "Sector Crawl");
+        let process_id = os.task_manager.get(task_id).expect("task should exist").process_id;
+
+        os.close_task(task_id).expect("close should succeed");
+
+        assert_eq!(
+            os.task_manager.get(task_id).expect("task should remain").state,
+            TaskState::Closed
+        );
+        assert_eq!(
+            os.process_manager.get(process_id).expect("process should remain").state,
+            ProcessState::Stopped
+        );
+    }
+
+    #[test]
+    fn crash_task_marks_task_and_process_crashed() {
+        let mut os = SystemOS::new(None);
+        let task_id = os.create_vm_game_task(42, "Sector Crawl");
+
+        os.crash_task(task_id, None).expect("crash should succeed");
+
+        assert_eq!(
+            os.task_manager.get(task_id).expect("task should exist").state,
+            TaskState::Crashed
+        );
+        assert_eq!(process_state_for_task(&os, task_id), ProcessState::Crashed);
+    }
+
+    #[test]
+    fn lifecycle_operation_missing_task_returns_typed_error() {
+        let mut os = SystemOS::new(None);
+
+        assert_eq!(os.suspend_task(TaskId(999)), Err(LifecycleError::TaskNotFound(TaskId(999))));
+    }
+
+    #[test]
+    fn lifecycle_operation_missing_process_returns_typed_error() {
+        let mut os = SystemOS::new(None);
+        let task_id = os.create_vm_game_task(42, "Sector Crawl");
+        let process_id = os.task_manager.get(task_id).expect("task should exist").process_id;
+
+        os.process_manager.mark_stopped(process_id);
+        os.process_manager.remove_stopped();
+
+        assert_eq!(os.suspend_task(task_id), Err(LifecycleError::ProcessNotFound(process_id)));
+    }
+
+    #[test]
+    fn resume_task_from_non_suspended_state_returns_invalid_transition() {
+        let mut os = SystemOS::new(None);
+        let task_id = os.create_vm_game_task(42, "Sector Crawl");
+
+        assert_eq!(
+            os.resume_task(task_id),
+            Err(LifecycleError::InvalidTransition {
+                task_id,
+                from: TaskState::Foreground,
+                operation: LifecycleOperation::Resume,
+            })
+        );
+    }
 }
diff --git a/discussion/index.ndjson b/discussion/index.ndjson
index 112b04fe..f2e2d13d 100644
--- a/discussion/index.ndjson
+++ b/discussion/index.ndjson
@@ -1,4 +1,4 @@
-{"type":"meta","next_id":{"DSC":34,"AGD":34,"DEC":25,"PLN":55,"LSN":41,"CLSN":1}}
+{"type":"meta","next_id":{"DSC":34,"AGD":34,"DEC":26,"PLN":58,"LSN":41,"CLSN":1}}
 {"type":"discussion","id":"DSC-0023","status":"done","ticket":"perf-full-migration-to-atomic-telemetry","title":"Agenda - [PERF] Full Migration to Atomic Telemetry","created_at":"2026-04-10","updated_at":"2026-04-10","tags":["perf","runtime","telemetry"],"agendas":[{"id":"AGD-0021","file":"AGD-0021-full-migration-to-atomic-telemetry.md","status":"done","created_at":"2026-04-10","updated_at":"2026-04-10"}],"decisions":[{"id":"DEC-0008","file":"DEC-0008-full-migration-to-atomic-telemetry.md","status":"accepted","created_at":"2026-04-10","updated_at":"2026-04-10"}],"plans":[{"id":"PLN-0007","file":"PLN-0007-full-migration-to-atomic-telemetry.md","status":"done","created_at":"2026-04-10","updated_at":"2026-04-10"}],"lessons":[{"id":"LSN-0028","file":"discussion/lessons/DSC-0023-perf-full-migration-to-atomic-telemetry/LSN-0028-converging-to-single-atomic-telemetry-source.md","status":"done","created_at":"2026-04-10","updated_at":"2026-04-10"}]}
 {"type":"discussion","id":"DSC-0020","status":"done","ticket":"jenkins-gitea-integration","title":"Jenkins Gitea Integration and Relocation","created_at":"2026-04-07","updated_at":"2026-04-07","tags":["ci","jenkins","gitea"],"agendas":[{"id":"AGD-0018","file":"AGD-0018-jenkins-gitea-integration-and-relocation.md","status":"done","created_at":"2026-04-07","updated_at":"2026-04-07"}],"decisions":[{"id":"DEC-0003","file":"DEC-0003-jenkins-gitea-strategy.md","status":"accepted","created_at":"2026-04-07","updated_at":"2026-04-07"}],"plans":[{"id":"PLN-0003","file":"PLN-0003-jenkins-gitea-execution.md","status":"done","created_at":"2026-04-07","updated_at":"2026-04-07"}],"lessons":[{"id":"LSN-0021","file":"discussion/lessons/DSC-0020-jenkins-gitea-integration/LSN-0021-jenkins-gitea-integration.md","status":"done","created_at":"2026-04-07","updated_at":"2026-04-07"}]}
 {"type":"discussion","id":"DSC-0021","status":"done","ticket":"asset-entry-codec-enum-with-metadata","title":"Asset Entry Codec Enum Contract","created_at":"2026-04-09","updated_at":"2026-04-09","tags":["asset","runtime","codec","metadata"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0024","file":"discussion/lessons/DSC-0021-asset-entry-codec-enum-contract/LSN-0024-string-on-the-wire-enum-in-runtime.md","status":"done","created_at":"2026-04-09","updated_at":"2026-04-09"}]}
@@ -30,5 +30,5 @@
 {"type":"discussion","id":"DSC-0019","status":"done","ticket":"jenkinsfile-correction","title":"Jenkinsfile Correction and Relocation","created_at":"2026-04-07","updated_at":"2026-04-07","tags":["ci","jenkins"],"agendas":[{"id":"AGD-0017","file":"AGD-0017-jenkinsfile-correction.md","status":"done","created_at":"2026-04-07","updated_at":"2026-04-07"}],"decisions":[{"id":"DEC-0002","file":"DEC-0002-jenkinsfile-strategy.md","status":"accepted","created_at":"2026-04-07","updated_at":"2026-04-07"}],"plans":[{"id":"PLN-0002","file":"PLN-0002-jenkinsfile-execution.md","status":"done","created_at":"2026-04-07","updated_at":"2026-04-07"}],"lessons":[{"id":"LSN-0020","file":"discussion/lessons/DSC-0019-jenkins-ci-standardization/LSN-0020-jenkins-standard-relocation.md","status":"done","created_at":"2026-04-07","updated_at":"2026-04-07"}]}
 {"type":"discussion","id":"DSC-0030","status":"done","ticket":"internal-viewport-270p","title":"Agenda - Internal Viewport 270p (480x270)","created_at":"2026-04-27","updated_at":"2026-04-28","tags":["gfx","runtime","viewport","resolution","frame-composer","host"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0039","file":"discussion/lessons/DSC-0030-internal-viewport-270p/LSN-0039-resolution-baseline-changes-are-runtime-wide-contracts.md","status":"done","created_at":"2026-04-28","updated_at":"2026-04-28"}]}
 {"type":"discussion","id":"DSC-0031","status":"done","ticket":"runtime-mode-separation-game-system","title":"Agenda - Runtime Mode Separation: Game and System","created_at":"2026-05-11","updated_at":"2026-05-14","tags":["runtime","firmware","hub","system-apps","game-mode","scheduler"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0040","file":"discussion/lessons/DSC-0031-runtime-mode-separation-game-system/LSN-0040-system-pipeline-separation.md","status":"done","created_at":"2026-05-14","updated_at":"2026-05-14"}]}
-{"type":"discussion","id":"DSC-0032","status":"open","ticket":"system-os-lifecycle-process-task-contract","title":"Agenda - SystemOS Lifecycle, Process and Task Contract","created_at":"2026-05-14","updated_at":"2026-05-14","tags":["runtime","os","lifecycle","process","task","shell","firmware"],"agendas":[{"id":"AGD-0032","file":"AGD-0032-system-os-lifecycle-process-task-contract.md","status":"open","created_at":"2026-05-14","updated_at":"2026-05-14"}],"decisions":[],"plans":[],"lessons":[]}
+{"type":"discussion","id":"DSC-0032","status":"in_progress","ticket":"system-os-lifecycle-process-task-contract","title":"Agenda - SystemOS Lifecycle, Process and Task Contract","created_at":"2026-05-14","updated_at":"2026-05-15","tags":["runtime","os","lifecycle","process","task","shell","firmware"],"agendas":[{"id":"AGD-0032","file":"AGD-0032-system-os-lifecycle-process-task-contract.md","status":"accepted","created_at":"2026-05-14","updated_at":"2026-05-15"}],"decisions":[{"id":"DEC-0025","file":"DEC-0025-systemos-lifecycle-authority-for-tasks-and-processes.md","status":"accepted","created_at":"2026-05-15","updated_at":"2026-05-15","ref_agenda":"AGD-0032"}],"plans":[{"id":"PLN-0055","file":"PLN-0055-systemos-lifecycle-core-api.md","status":"done","created_at":"2026-05-15","updated_at":"2026-05-15","ref_decisions":["DEC-0025"]},{"id":"PLN-0056","file":"PLN-0056-firmware-lifecycle-delegation.md","status":"done","created_at":"2026-05-15","updated_at":"2026-05-15","ref_decisions":["DEC-0025"]},{"id":"PLN-0057","file":"PLN-0057-lifecycle-tests-and-invariants.md","status":"done","created_at":"2026-05-15","updated_at":"2026-05-15","ref_decisions":["DEC-0025"]}],"lessons":[]}
 {"type":"discussion","id":"DSC-0033","status":"in_progress","ticket":"system-os-service-ownership-and-module-layout","title":"Agenda - SystemOS Service Ownership and Module Layout","created_at":"2026-05-14","updated_at":"2026-05-14","tags":["runtime","os","services","module-layout","vm","window-manager","logging"],"agendas":[{"id":"AGD-0033","file":"AGD-0033-system-os-service-ownership-and-module-layout.md","status":"accepted","created_at":"2026-05-14","updated_at":"2026-05-14"}],"decisions":[{"id":"DEC-0024","file":"DEC-0024-systemos-service-ownership-and-module-layout.md","status":"accepted","created_at":"2026-05-14","updated_at":"2026-05-14","ref_agenda":"AGD-0033"}],"plans":[{"id":"PLN-0051","file":"PLN-0051-move-vm-runtime-into-services.md","status":"done","created_at":"2026-05-14","updated_at":"2026-05-14","ref_decisions":["DEC-0024"]},{"id":"PLN-0052","file":"PLN-0052-promote-window-manager-to-systemos-service.md","status":"done","created_at":"2026-05-14","updated_at":"2026-05-14","ref_decisions":["DEC-0024"]},{"id":"PLN-0053","file":"PLN-0053-move-logging-ownership-to-systemos.md","status":"done","created_at":"2026-05-14","updated_at":"2026-05-14","ref_decisions":["DEC-0024"]},{"id":"PLN-0054","file":"PLN-0054-move-fs-and-memcard-ownership-to-systemos.md","status":"done","created_at":"2026-05-14","updated_at":"2026-05-14","ref_decisions":["DEC-0024"]}],"lessons":[]}
diff --git a/discussion/workflow/agendas/AGD-0032-system-os-lifecycle-process-task-contract.md b/discussion/workflow/agendas/AGD-0032-system-os-lifecycle-process-task-contract.md
index 425b59cb..913f6184 100644
--- a/discussion/workflow/agendas/AGD-0032-system-os-lifecycle-process-task-contract.md
+++ b/discussion/workflow/agendas/AGD-0032-system-os-lifecycle-process-task-contract.md
@@ -2,10 +2,10 @@
 id: AGD-0032
 ticket: system-os-lifecycle-process-task-contract
 title: Agenda - SystemOS Lifecycle, Process and Task Contract
-status: open
+status: accepted
 created: 2026-05-14
-resolved:
-decision:
+resolved: 2026-05-15
+decision: DEC-0025
 tags: [runtime, os, lifecycle, process, task, shell, firmware]
 ---
 
@@ -65,7 +65,7 @@ AppMode::Game
     -> marca task como Foreground
     -> entra em GameRunningStep com TaskId
 
-AppMode::System
+AppMode::Shell
     -> cria processo VmShell
     -> cria task Shell
     -> marca task como Foreground
@@ -249,38 +249,83 @@ O primeiro corte deve ser pequeno:
 O `WindowManager` deve ser a agenda/decisão seguinte: promovê-lo a serviço real
 do sistema quando a base `Task/Process/Lifecycle` estiver fechada.
 
-Também precisamos resolver a nomenclatura `System` versus `Shell`. O código e o
-contrato anterior usam `AppMode::System`; a narrativa nova fala em
-`AppMode::Shell`. A agenda deve decidir se:
-
-- mantemos `AppMode::System` como perfil formal e usamos `Shell` como tipo de
-  task/processo/UI; ou
-- renomeamos o perfil público para `Shell`, com migração explícita.
-
-Minha inclinação inicial é manter `AppMode::System` por compatibilidade de
-contrato e usar `Shell` para a camada de UI/app gerenciado dentro do OS.
+`AppMode::System` já foi renomeado para `AppMode::Shell`. A nomenclatura pública
+fica alinhada com o modelo atual: `AppMode::Game` executa cartuchos como
+jogo/session fullscreen, enquanto `AppMode::Shell` executa cartuchos como apps
+gerenciados pela Shell do Prometeu OS.
 
 ## Perguntas em Aberto
 
-- [ ] `SystemOS` deve expor `suspend_task`, `resume_task`, `close_task` e
-  `crash_task` já no próximo corte?
-- [ ] `foreground_task` e `background_task` devem ser operações explícitas do
-  `SystemOS`, ou apenas detalhe do `TaskManager`?
-- [ ] Qual é a semântica exata de `Background`: processo ainda `Running`,
+- [x] `SystemOS` deve expor `suspend_task`, `resume_task`, `close_task` e
+  `crash_task` já no próximo corte? Sim. Eles são a próxima API semântica de
+  lifecycle e evitam que firmware ou outros chamadores manipulem
+  `TaskManager` e `ProcessManager` separadamente.
+- [x] `foreground_task` e `background_task` devem ser operações explícitas do
+  `SystemOS`, ou apenas detalhe do `TaskManager`? `set_foreground_task(task_id)`
+  deve ser operação explícita do `SystemOS`. `background_task` fica fora da
+  primeira wave.
+- [x] Qual é a semântica exata de `Background`: processo ainda `Running`,
   processo `Suspended`, ou estado de presença sem garantia de execução?
-- [ ] `resume_task` deve sempre mover a task para `Foreground`, ou pode retomar
-  para `Background`?
-- [ ] O que acontece quando a task existe mas o processo associado não existe?
-- [ ] `close_task` remove imediatamente entidades fechadas ou apenas marca
-  `Closed`/`Stopped` para coleta posterior?
-- [ ] `crash_task` deve receber/armazenar `CrashReport`, ou isso fica em outro
-  serviço de diagnóstico?
-- [ ] `GameRunningStep` deve transicionar para Hub/Home quando sua task deixa de
-  estar `Foreground`, ou apenas não executar frame?
-- [ ] O perfil de manifesto continua `AppMode::System`, ou abrimos uma migração
-  formal para `AppMode::Shell`?
-- [ ] A decisão de ownership/layout em `AGD-0033` deve ser pré-requisito para o
-  primeiro plan de lifecycle, ou pode ser executada em paralelo?
+  Fora de escopo nesta wave. `Background` permanece estado reservado sem
+  semântica normativa no primeiro contrato.
+- [x] `resume_task` deve sempre mover a task para `Foreground`, ou pode retomar
+  para `Background`? Nesta wave, `resume_task` sempre move para
+  `Foreground` e move o processo associado para `Running`.
+- [x] O que acontece quando a task existe mas o processo associado não existe?
+  Isso é erro de lifecycle e deve retornar erro tipado, não `bool`.
+- [x] `close_task` remove imediatamente entidades fechadas ou apenas marca
+  `Closed`/`Stopped` para coleta posterior? Apenas marca `TaskState::Closed` e
+  `ProcessState::Stopped`; coleta é operação futura separada.
+- [x] `crash_task` deve receber/armazenar `CrashReport`, ou isso fica em outro
+  serviço de diagnóstico? A operação deve aceitar ou preparar o contrato para
+  `CrashReport`, mas a primeira wave pode apenas coordenar estados e deixar o
+  armazenamento para diagnóstico/logging futuro.
+- [x] `GameRunningStep` deve transicionar para Hub/Home quando sua task deixa de
+  estar `Foreground`, ou apenas não executar frame? Nesta wave, mantém a
+  invariant forte: `GameRunningStep` só suporta game task em `Foreground`; outro
+  estado é inválido e deve gerar diagnóstico/crash claro.
+- [x] O perfil de manifesto continua `AppMode::System`, ou abrimos uma migração
+  formal para `AppMode::Shell`? Resolvido: `AppMode::System` já virou
+  `AppMode::Shell`.
+- [x] A decisão de ownership/layout em `AGD-0033` deve ser pré-requisito para o
+  primeiro plan de lifecycle, ou pode ser executada em paralelo? Pode executar
+  em paralelo e não bloqueia esta wave. O plan de lifecycle deve evitar mover
+  `WindowManager`, `VirtualMachineRuntime`, Hub ou background services.
+
+## Resolucao
+
+A agenda fecha pela **Opcao B**: centralizar lifecycle no `SystemOS`.
+
+`SystemOS` deixa de ser apenas um agregador de serviços e passa a ser a
+autoridade normativa para coordenar `TaskState` e `ProcessState`. `TaskManager`
+e `ProcessManager` continuam existindo, mas como mecanismos internos de
+armazenamento e transição simples. O firmware não deve decidir diretamente como
+task e processo mudam juntos.
+
+A primeira wave deve cobrir apenas:
+
+- `set_foreground_task(task_id)`
+- `suspend_task(task_id)`
+- `resume_task(task_id)`
+- `close_task(task_id)`
+- `crash_task(task_id, report?)`
+
+O mapeamento inicial é:
+
+```text
+TaskState::Foreground -> ProcessState::Running
+TaskState::Suspended  -> ProcessState::Suspended
+TaskState::Closed     -> ProcessState::Stopped
+TaskState::Crashed    -> ProcessState::Crashed
+```
+
+`Background` permanece reservado e explicitamente fora da primeira wave.
+
+O primeiro plan recomendado é `PLAN-0032-A: SystemOS Lifecycle API`, com escopo
+limitado a erro tipado, métodos de lifecycle em `SystemOS`, coordenação de
+`TaskState`/`ProcessState`, testes unitários no nível `SystemOS` e nenhuma
+mudança em `WindowManager`, `PrometeuHub`, background services ou UX completa de
+suspensão/retomada.
 
 ## Criterio para Encerrar
 
diff --git a/discussion/workflow/decisions/DEC-0025-systemos-lifecycle-authority-for-tasks-and-processes.md b/discussion/workflow/decisions/DEC-0025-systemos-lifecycle-authority-for-tasks-and-processes.md
new file mode 100644
index 00000000..de7fe393
--- /dev/null
+++ b/discussion/workflow/decisions/DEC-0025-systemos-lifecycle-authority-for-tasks-and-processes.md
@@ -0,0 +1,203 @@
+---
+id: DEC-0025
+ticket: system-os-lifecycle-process-task-contract
+title: SystemOS Lifecycle Authority for Tasks and Processes
+status: accepted
+created: 2026-05-15
+ref_agenda: AGD-0032
+tags: [runtime, os, lifecycle, process, task, shell, firmware]
+---
+
+## Status
+
+Accepted. Esta decisão foi aceita em 2026-05-15 e passa a ser referência
+normativa para a primeira wave de lifecycle de `Task` e `Process` no
+`SystemOS`.
+
+## Contexto
+
+`SystemOS` já agrega serviços centrais do Prometeu OS, incluindo runtime VM,
+processos e tasks. A arquitetura anterior ainda permitia que firmware,
+`TaskManager` e `ProcessManager` decidissem transições de lifecycle de forma
+distribuída.
+
+Esse arranjo é insuficiente para um console OS: `Task` representa presença
+navegável e experiência do usuário, enquanto `Process` representa execução
+técnica. Essas duas dimensões precisam mudar juntas por uma autoridade
+semântica única.
+
+`DEC-0024` já separou ownership/layout de serviços do `SystemOS`. Esta decisão
+é focada apenas em lifecycle de `Task`/`Process` e não reabre a migração de
+serviços, `WindowManager`, Hub ou ownership da VM.
+
+`AppMode::System` já foi renomeado para `AppMode::Shell`. O contrato público
+fica:
+
+- `AppMode::Game`: cartucho executado como jogo/session fullscreen.
+- `AppMode::Shell`: cartucho executado como app gerenciado pela Shell do
+  Prometeu OS.
+
+## Decisao
+
+`SystemOS` SHALL ser a autoridade normativa de lifecycle para `Task` e
+`Process`.
+
+`TaskManager` e `ProcessManager` SHALL continuar existindo como mecanismos
+internos de armazenamento e transição local, mas MUST NOT ser a API semântica
+principal para lifecycle coordenado. Firmware e outros chamadores MUST delegate
+transições semânticas de lifecycle para `SystemOS`.
+
+A primeira wave de lifecycle SHALL expor o contrato mínimo:
+
+```text
+set_foreground_task(task_id)
+suspend_task(task_id)
+resume_task(task_id)
+close_task(task_id)
+crash_task(task_id, report?)
+```
+
+Essas operações SHALL coordenar `TaskState` e `ProcessState` juntos.
+
+O mapeamento normativo inicial SHALL ser:
+
+```text
+TaskState::Foreground -> ProcessState::Running
+TaskState::Suspended  -> ProcessState::Suspended
+TaskState::Closed     -> ProcessState::Stopped
+TaskState::Crashed    -> ProcessState::Crashed
+```
+
+`Background` SHALL remain reserved and MUST NOT guide any public lifecycle API
+in the first wave.
+
+`resume_task` SHALL mean resuming active user-perceived execution. In this first
+wave, it MUST move a suspended task to `TaskState::Foreground` and its associated
+process to `ProcessState::Running`.
+
+`close_task` SHALL mark task/process state only. It MUST NOT remove task or
+process entities immediately. Collection of closed/stopped entities is a future
+separate operation.
+
+`crash_task` SHALL be the semantic crash operation for a task. The first wave MAY
+only coordinate `TaskState::Crashed` and `ProcessState::Crashed`, but the
+signature SHOULD either accept or remain compatible with a future `CrashReport`
+handoff to diagnostics/logging.
+
+`GameRunningStep` SHALL keep the strong invariant that an active game execution
+step only supports a game task in `TaskState::Foreground`. In this wave, any
+non-foreground game task observed by `GameRunningStep` is invalid and should
+produce clear diagnostic/crash behavior rather than silently running frames.
+
+## Rationale
+
+Centralizing lifecycle in `SystemOS` prevents firmware from becoming the de
+facto OS policy layer. Firmware should remain a boot/execution state machine; it
+should not know how to update `TaskState` and `ProcessState` in lockstep.
+
+Keeping `TaskManager` and `ProcessManager` as internal mechanisms preserves
+simple storage and local transition responsibilities while giving the OS a
+single semantic boundary for suspend, resume, close, crash and foreground
+selection.
+
+Leaving `Background` out of the first wave avoids freezing an incomplete model
+for docked apps, background services, non-focused shell apps or capability-gated
+execution. The reserved enum state can remain, but no normative public lifecycle
+API should depend on it yet.
+
+Making `resume_task` return to `Foreground` keeps the first contract simple and
+matches the user-perceived meaning of resume. Future background wake/service
+semantics should be introduced as separate concepts rather than overloading
+`resume_task`.
+
+Marking closed/stopped entities instead of removing them immediately preserves
+room for diagnostics, crash handling, system UI history and later collection
+policy.
+
+## Invariantes / Contrato
+
+- `SystemOS` MUST coordinate lifecycle operations that affect both task and
+  process state.
+- Firmware MUST NOT call separate task/process manager transitions to implement
+  semantic lifecycle operations such as suspend, resume, close or crash.
+- `set_foreground_task(task_id)` MUST set the task to `TaskState::Foreground`
+  and its associated process to `ProcessState::Running`.
+- `suspend_task(task_id)` MUST set the task to `TaskState::Suspended` and its
+  associated process to `ProcessState::Suspended`.
+- `resume_task(task_id)` MUST move from suspended user-visible execution back to
+  foreground execution: `TaskState::Foreground` and `ProcessState::Running`.
+- `close_task(task_id)` MUST set `TaskState::Closed` and
+  `ProcessState::Stopped`.
+- `crash_task(task_id, report?)` MUST set `TaskState::Crashed` and
+  `ProcessState::Crashed`.
+- A missing task MUST return a typed lifecycle error such as
+  `LifecycleError::TaskNotFound(TaskId)`.
+- A task whose associated process is missing MUST return a typed lifecycle error
+  such as `LifecycleError::ProcessNotFound(ProcessId)`.
+- Invalid lifecycle transitions SHOULD return a typed error carrying the task,
+  source state and requested operation.
+- `Background` MUST NOT have normative first-wave semantics.
+- This decision MUST NOT move `WindowManager`, redesign `PrometeuHub`, introduce
+  background services, implement docking, implement app switching, or make
+  `ProcessManager` own the VM.
+
+The expected first-wave error shape is:
+
+```rust
+pub enum LifecycleError {
+    TaskNotFound(TaskId),
+    ProcessNotFound(ProcessId),
+    InvalidTransition {
+        task_id: TaskId,
+        from: TaskState,
+        operation: LifecycleOperation,
+    },
+}
+```
+
+`LifecycleOperation` SHOULD identify at least:
+
+```text
+SetForeground
+Suspend
+Resume
+Close
+Crash
+```
+
+## Impactos
+
+- `prometeu-system` must expose lifecycle methods on `SystemOS`.
+- `TaskManager` and `ProcessManager` may need smaller internal transition
+  helpers, but not public semantic ownership of lifecycle.
+- Firmware code should call `system_os.suspend_task(...)`,
+  `system_os.resume_task(...)`, `system_os.close_task(...)`,
+  `system_os.crash_task(...)` and `system_os.set_foreground_task(...)` instead
+  of coordinating task/process managers directly.
+- Tests should move lifecycle assertions to the `SystemOS` level.
+- Specs do not need to change before the first code plan unless the lifecycle
+  API is promoted to a public external contract.
+- `DEC-0024` remains valid and complementary; ownership/layout work can execute
+  in parallel and does not block this first lifecycle wave.
+
+## Referencias
+
+- Agenda: `AGD-0032`
+- Related decision: `DEC-0024`
+
+## Propagacao Necessaria
+
+- Plan: create `PLAN-0032-A: SystemOS Lifecycle API`.
+- Code: update `prometeu-system` `SystemOS`, task/process lifecycle helpers,
+  firmware callsites if they currently coordinate managers directly, and unit
+  tests.
+- Tests: add unit coverage for successful lifecycle mappings, missing task,
+  missing process and invalid transition handling.
+- Discussion: keep `Background` reserved for a future agenda covering dock,
+  services, non-focused shell apps and capability-gated execution.
+- Learn: after implementation is complete, write or update lessons only from the
+  published state.
+
+## Revisao
+
+- 2026-05-15: Initial accepted decision from `AGD-0032`.
diff --git a/discussion/workflow/plans/PLN-0055-systemos-lifecycle-core-api.md b/discussion/workflow/plans/PLN-0055-systemos-lifecycle-core-api.md
new file mode 100644
index 00000000..51bfd989
--- /dev/null
+++ b/discussion/workflow/plans/PLN-0055-systemos-lifecycle-core-api.md
@@ -0,0 +1,121 @@
+---
+id: PLN-0055
+ticket: system-os-lifecycle-process-task-contract
+title: SystemOS Lifecycle Core API
+status: done
+created: 2026-05-15
+ref_decisions: [DEC-0025]
+tags: [runtime, os, lifecycle, process, task, shell, firmware]
+---
+
+## Briefing
+
+Implement the first-wave `SystemOS` lifecycle API required by `DEC-0025`.
+`SystemOS` becomes the semantic coordinator for `TaskState` and `ProcessState`;
+`TaskManager` and `ProcessManager` remain storage and simple transition
+mechanisms.
+
+## Source Decisions
+
+- `DEC-0025`: `SystemOS` is the lifecycle authority for tasks and processes.
+
+## Target
+
+Add typed lifecycle operations to `prometeu-system` without changing firmware
+behavior, WindowManager ownership, Hub UI, background services, docking, app
+switching, or VM ownership.
+
+## Scope
+
+- Add `LifecycleError` and `LifecycleOperation` in `prometeu-system`.
+- Add lifecycle methods on `SystemOS`:
+  - `set_foreground_task(task_id)`
+  - `suspend_task(task_id)`
+  - `resume_task(task_id)`
+  - `close_task(task_id)`
+  - `crash_task(task_id, report?)`
+- Coordinate task and process state changes inside `SystemOS`.
+- Preserve `TaskManager` and `ProcessManager` as internal state managers.
+- Keep `Background` reserved and out of public lifecycle semantics.
+
+## Out of Scope
+
+- Background semantics.
+- Docking.
+- Background services.
+- App switcher.
+- WindowManager migration.
+- PrometeuHub redesign.
+- Game suspend/resume UX.
+- ProcessManager owning VM execution.
+- Collection/removal of closed or stopped entities.
+
+## Execution Plan
+
+1. Add lifecycle types.
+   - Create a lifecycle module or colocate the types under
+     `crates/console/prometeu-system/src/os/`.
+   - Define `LifecycleOperation` with `SetForeground`, `Suspend`, `Resume`,
+     `Close`, and `Crash`.
+   - Define `LifecycleError` with `TaskNotFound(TaskId)`,
+     `ProcessNotFound(ProcessId)`, and `InvalidTransition { task_id, from,
+     operation }`.
+   - Re-export the types through `prometeu-system` only if the existing public
+     module style requires it.
+
+2. Add the `SystemOS` coordination helpers.
+   - Target file: `crates/console/prometeu-system/src/os/system_os.rs`.
+   - Resolve the task first through `task_manager`.
+   - Extract the task's associated `ProcessId`.
+   - Validate that the process exists before mutating state.
+   - Return typed errors instead of `bool`.
+
+3. Implement first-wave operations.
+   - `set_foreground_task`: call `task_manager.set_foreground(task_id)` and
+     `process_manager.mark_running(process_id)`.
+   - `suspend_task`: set `TaskState::Suspended` and
+     `ProcessState::Suspended`.
+   - `resume_task`: require a suspended task, then set
+     `TaskState::Foreground` and `ProcessState::Running`.
+   - `close_task`: set `TaskState::Closed` and `ProcessState::Stopped`.
+   - `crash_task`: set `TaskState::Crashed` and `ProcessState::Crashed`.
+
+4. Keep manager APIs mechanical.
+   - Target files:
+     `crates/console/prometeu-system/src/services/task/task_manager.rs` and
+     `crates/console/prometeu-system/src/services/process/process_manager.rs`.
+   - Do not remove existing manager methods in this plan.
+   - Add only narrow helpers if needed to support the `SystemOS` API cleanly.
+
+5. Update task creation to use the lifecycle API internally.
+   - In `SystemOS::create_vm_game_task`, `create_vm_shell_task`, and
+     `create_native_shell_task`, replace direct foreground manager calls with
+     `set_foreground_task`.
+   - Keep the constructors returning `TaskId`; handle impossible internal
+     errors with explicit debug/assert behavior or an internal helper that
+     cannot fail after creation.
+
+## Acceptance Criteria
+
+- `SystemOS` exposes all five first-wave lifecycle methods from `DEC-0025`.
+- All five methods coordinate both task and process state.
+- Missing task returns `LifecycleError::TaskNotFound`.
+- Missing associated process returns `LifecycleError::ProcessNotFound`.
+- Invalid resume from a non-suspended task returns `InvalidTransition`.
+- `Background` is not exposed through a new lifecycle API.
+- Closed and stopped entities are not removed by `close_task`.
+
+## Tests / Validation
+
+- Add or update unit tests in `prometeu-system` for each lifecycle mapping.
+- Add tests for missing task, missing process, and invalid resume transition.
+- Run the affected Rust test suite for `prometeu-system`.
+
+## Risks
+
+- Existing direct manager access is public enough that callsites may keep using
+  it. This plan does not remove those methods; enforcement continues in
+  `PLN-0056`.
+- `TaskManager::set_foreground` currently moves the previous foreground task to
+  `Background`. Since `Background` is out of first-wave semantics, tests must
+  avoid treating that transition as normative lifecycle policy.
diff --git a/discussion/workflow/plans/PLN-0056-firmware-lifecycle-delegation.md b/discussion/workflow/plans/PLN-0056-firmware-lifecycle-delegation.md
new file mode 100644
index 00000000..5c2500d9
--- /dev/null
+++ b/discussion/workflow/plans/PLN-0056-firmware-lifecycle-delegation.md
@@ -0,0 +1,119 @@
+---
+id: PLN-0056
+ticket: system-os-lifecycle-process-task-contract
+title: Firmware Lifecycle Delegation
+status: done
+created: 2026-05-15
+ref_decisions: [DEC-0025]
+tags: [runtime, os, lifecycle, process, task, shell, firmware]
+---
+
+## Briefing
+
+Move firmware lifecycle callsites to the `SystemOS` semantic API introduced by
+`PLN-0055`. Firmware remains a boot and execution state machine; it must not
+coordinate task/process lifecycle by calling `TaskManager` and `ProcessManager`
+separately.
+
+## Source Decisions
+
+- `DEC-0025`: firmware must delegate semantic lifecycle transitions to
+  `SystemOS`.
+
+## Target
+
+Replace direct lifecycle manipulation in firmware with `SystemOS` lifecycle
+methods while preserving current observable behavior.
+
+## Scope
+
+- Update cartridge launch paths to use `SystemOS` lifecycle authority when
+  placing newly created tasks in foreground.
+- Update crash paths where a task crash is detected so task/process state moves
+  through `SystemOS::crash_task`.
+- Preserve the strong `GameRunningStep` invariant: a running game step only
+  supports a game task in `TaskState::Foreground`.
+- Keep `SystemRunningStep`, Hub, and WindowManager behavior unchanged except for
+  lifecycle delegation required by `DEC-0025`.
+
+## Out of Scope
+
+- Full game suspend/resume UX.
+- Returning to Hub/Home on suspend or close.
+- Shell UI redesign.
+- App switcher.
+- Background task API.
+- WindowManager migration.
+- Diagnostics service or crash report persistence.
+
+## Execution Plan
+
+1. Audit direct lifecycle manager usage.
+   - Target files:
+     `crates/console/prometeu-firmware/src/firmware/*.rs`,
+     `crates/console/prometeu-system/src/os/system_os.rs`.
+   - Search for direct calls to `task_manager.set_foreground`,
+     `task_manager.mark_suspended`, `task_manager.close_task`,
+     `task_manager.mark_crashed`, `process_manager.mark_suspended`,
+     `process_manager.mark_stopped`, and `process_manager.mark_crashed`.
+
+2. Keep launch construction inside `SystemOS`.
+   - Target file:
+     `crates/console/prometeu-system/src/os/system_os.rs`.
+   - Ensure `create_vm_game_task`, `create_vm_shell_task`, and
+     `create_native_shell_task` coordinate foreground/running state through the
+     lifecycle core API from `PLN-0055`.
+
+3. Update game crash handling.
+   - Target file:
+     `crates/console/prometeu-firmware/src/firmware/firmware_step_game_running.rs`.
+   - When the task is missing, preserve current crash-screen behavior. If the
+     task does not exist, `SystemOS::crash_task` cannot attach state and the
+     firmware should continue producing a clear crash report.
+   - When a foreground invariant fails for an existing task, call
+     `ctx.os.crash_task(task_id, report?)` before transitioning to
+     `AppCrashesStep`.
+   - When `tick_vm` returns a crash report for the active task, call
+     `ctx.os.crash_task(task_id, report?)` before transitioning to
+     `AppCrashesStep`.
+
+4. Keep shell launch behavior unchanged.
+   - Target file:
+     `crates/console/prometeu-firmware/src/firmware/firmware_step_load_cartridge.rs`.
+   - Do not move `WindowManager` logic.
+   - Do not change shell window creation or focus behavior.
+   - Rely on `SystemOS::create_vm_shell_task` to coordinate task/process
+     foreground state.
+
+5. Avoid introducing background lifecycle.
+   - Do not add `background_task`.
+   - Do not treat `TaskState::Background` as a public lifecycle result in
+     firmware.
+
+## Acceptance Criteria
+
+- Firmware does not coordinate task/process lifecycle by manually changing both
+  managers.
+- Game VM crash results in `SystemOS::crash_task` for the active task when the
+  task still exists.
+- Foreground invariant violation in `GameRunningStep` crashes the task through
+  `SystemOS`.
+- Launch behavior for `AppMode::Game` and `AppMode::Shell` remains unchanged.
+- WindowManager and Hub behavior are not redesigned or moved.
+
+## Tests / Validation
+
+- Run firmware tests that cover cartridge launch for `AppMode::Game` and
+  `AppMode::Shell`.
+- Add or update tests for `GameRunningStep` crash delegation if the existing
+  test harness supports firmware step testing.
+- Run the affected Rust test suite for `prometeu-firmware`.
+
+## Risks
+
+- Crash report ownership is not final. This plan should pass the report through
+  the lifecycle API only to the extent supported by `PLN-0055`; persistence
+  remains out of scope.
+- `GameRunningStep` currently reads task state directly. This is acceptable for
+  enforcing the foreground invariant, but it must not become a place that
+  coordinates lifecycle state across managers.
diff --git a/discussion/workflow/plans/PLN-0057-lifecycle-tests-and-invariants.md b/discussion/workflow/plans/PLN-0057-lifecycle-tests-and-invariants.md
new file mode 100644
index 00000000..99a4b181
--- /dev/null
+++ b/discussion/workflow/plans/PLN-0057-lifecycle-tests-and-invariants.md
@@ -0,0 +1,110 @@
+---
+id: PLN-0057
+ticket: system-os-lifecycle-process-task-contract
+title: Lifecycle Tests and Invariants
+status: done
+created: 2026-05-15
+ref_decisions: [DEC-0025]
+tags: [runtime, os, lifecycle, process, task, shell, firmware]
+---
+
+## Briefing
+
+Add focused test coverage and invariant checks for the lifecycle contract
+introduced by `PLN-0055` and consumed by `PLN-0056`.
+
+## Source Decisions
+
+- `DEC-0025`: first-wave lifecycle maps task states to process states through
+  `SystemOS` and keeps `Background` outside the public lifecycle contract.
+
+## Target
+
+Make the lifecycle contract executable through unit tests and regression tests,
+without expanding scope into background execution, UI behavior, or service
+ownership.
+
+## Scope
+
+- Add `SystemOS` unit tests for all first-wave lifecycle operations.
+- Add negative tests for missing task, missing process, and invalid transition.
+- Add firmware-level regression coverage for lifecycle delegation where the
+  existing test harness can observe it.
+- Keep manager tests focused on local storage and simple transitions.
+
+## Out of Scope
+
+- End-to-end UX tests for suspend/resume.
+- Hub/Home navigation after lifecycle changes.
+- Background services.
+- Docking.
+- App switching.
+- Diagnostics persistence.
+- WindowManager behavior changes.
+
+## Execution Plan
+
+1. Add core lifecycle mapping tests.
+   - Target file: `crates/console/prometeu-system/src/os/system_os.rs` or a
+     dedicated `os` test module following existing crate style.
+   - Verify:
+     - `set_foreground_task`: task becomes `Foreground`, process becomes
+       `Running`.
+     - `suspend_task`: task becomes `Suspended`, process becomes `Suspended`.
+     - `resume_task`: suspended task becomes `Foreground`, process becomes
+       `Running`.
+     - `close_task`: task becomes `Closed`, process becomes `Stopped`.
+     - `crash_task`: task becomes `Crashed`, process becomes `Crashed`.
+
+2. Add error tests.
+   - Verify missing task returns `LifecycleError::TaskNotFound`.
+   - Verify a task whose process no longer exists returns
+     `LifecycleError::ProcessNotFound`.
+   - Verify invalid resume from non-suspended task returns
+     `LifecycleError::InvalidTransition` with operation `Resume`.
+
+3. Add non-removal tests.
+   - Verify `close_task` leaves the closed task available through
+     `TaskManager::get`.
+   - Verify `close_task` leaves the stopped process available through
+     `ProcessManager::get`.
+   - Do not call `remove_closed` or `remove_stopped` as part of lifecycle API
+     tests.
+
+4. Add background exclusion tests.
+   - Assert there is no `SystemOS::background_task` API added by this wave.
+   - Keep any existing `TaskState::Background` behavior in manager-level tests
+     documented as manager behavior, not first-wave lifecycle semantics.
+
+5. Add firmware regression tests after `PLN-0056`.
+   - Target file: `crates/console/prometeu-firmware/src/firmware/firmware.rs`
+     or the relevant step test module.
+   - Verify game launch still enters `GameRunningStep`.
+   - Verify shell launch still enters `SystemRunningStep`.
+   - Verify a game crash path marks the active task as crashed when the test
+     harness can observe `ctx.os.task_manager`.
+
+## Acceptance Criteria
+
+- Every lifecycle method from `DEC-0025` has a positive `SystemOS` test.
+- Missing task and missing process return typed lifecycle errors.
+- Invalid resume transition is covered by a test.
+- `close_task` is tested as mark-only, not removal.
+- No test encodes `Background` as a normative first-wave lifecycle operation.
+- Firmware regression tests still pass for game and shell launch.
+
+## Tests / Validation
+
+- Run `cargo test -p prometeu-system`.
+- Run `cargo test -p prometeu-firmware`.
+- If workspace-level tests are practical after these changes, run the relevant
+  workspace command used by this repository.
+
+## Risks
+
+- Some negative tests may require test-only helpers to create an inconsistent
+  task/process relationship. Prefer narrow test fixtures over making corruption
+  APIs public.
+- Compile-time absence of `background_task` may not be directly testable
+  without compile-fail infrastructure. In that case, validate it by code review
+  and keep runtime tests focused on the exposed first-wave methods.