fix: model-loaded guards on transcribe_pcm, power assertion guards on tasks LLM commands

transcribe_pcm and transcribe_pcm_parakeet did not check that their
respective engines were loaded before clone+spawn_blocking. transcribe_file
already calls ensure_model_loaded; these now mirror that posture with a
friendly error when the engine is unloaded, matching what
extract_content_tags_cmd does.

decompose_and_store and extract_tasks_from_transcript_cmd ran multi-second
LLM inference inside spawn_blocking without the PowerAssertion guard that
cleanup_transcript_text_cmd and extract_content_tags_cmd already use. Both
now begin a guard so the macOS App-Nap inhibitor (and the planned
Linux/Windows equivalents per KI-02, KI-03) can pin the process for the
duration of the inference.

HANDOVER.md gets a status note clarifying it captures Phase 9 state. The
schema head referenced (v14) was the head at session time; current head
is v15 (`idx_transcripts_profile_created` composite index) per the
architecture map.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src-tauri/src/commands/tasks.rs

@@ -18,6 +18,7 @@ use magnotia_storage::{
     FeedbackRow, FeedbackTargetType, TaskRow,
 };
 
+use crate::commands::power::PowerAssertion;
 use crate::AppState;
 
 /// Frontend-facing task shape. Matches the in-memory object in page.svelte.js.
@@ -319,6 +320,7 @@ pub async fn decompose_and_store(
     let engine = state.llm_engine.clone();
     let parent_text = parent.text.clone();
     let steps = tokio::task::spawn_blocking(move || {
+        let _power_guard = PowerAssertion::begin("magnotia LLM micro-step decomposition");
         engine.decompose_task_with_feedback(&parent_text, &examples)
     })
     .await
@@ -361,10 +363,13 @@ pub async fn extract_tasks_from_transcript_cmd(
     .unwrap_or_default();
 
     let engine = state.llm_engine.clone();
-    tokio::task::spawn_blocking(move || engine.extract_tasks_with_feedback(&transcript, &examples))
-        .await
-        .map_err(|e| e.to_string())?
-        .map_err(|e| e.to_string())
+    tokio::task::spawn_blocking(move || {
+        let _power_guard = PowerAssertion::begin("magnotia LLM task extraction");
+        engine.extract_tasks_with_feedback(&transcript, &examples)
+    })
+    .await
+    .map_err(|e| e.to_string())?
+    .map_err(|e| e.to_string())
 }
 
 #[tauri::command]

src-tauri/src/commands/transcription.rs

@@ -154,6 +154,11 @@ pub async fn transcribe_pcm(
     profile_id: Option<String>,
 ) -> Result<(), String> {
     ensure_main_window(&window)?;
+    if !state.whisper_engine.is_loaded() {
+        return Err(
+            "Whisper model not loaded. Download a Whisper model in Settings.".to_string(),
+        );
+    }
     let resolved_profile_id =
         profile_id.unwrap_or_else(|| magnotia_storage::DEFAULT_PROFILE_ID.to_string());
 
@@ -352,6 +357,11 @@ pub async fn transcribe_pcm_parakeet(
     profile_id: Option<String>,
 ) -> Result<(), String> {
     ensure_main_window(&window)?;
+    if !state.parakeet_engine.is_loaded() {
+        return Err(
+            "Parakeet model not loaded. Enable Parakeet in Settings.".to_string(),
+        );
+    }
     let resolved_profile_id =
         profile_id.unwrap_or_else(|| magnotia_storage::DEFAULT_PROFILE_ID.to_string());