docs: update backlog tasks

Author: pythoninthegrass

…rebaked-image-macOS-runner-broadening.md …rebaked-image-macOS-runner-broadening.mdbacklog/tasks/task-306 - CI-Build-Time-Iteration-Linux-prebaked-image-macOS-runner-broadening.md renamed to backlog/completed/task-306 - CI-Build-Time-Iteration-Linux-prebaked-image-macOS-runner-broadening.md backlog/tasks/task-306 - CI-Build-Time-Iteration-Linux-prebaked-image-macOS-runner-broadening.md renamed to backlog/completed/task-306 - CI-Build-Time-Iteration-Linux-prebaked-image-macOS-runner-broadening.md

@@ -0,0 +1,43 @@
+---
+id: TASK-310
+title: Add lyrics-based validation to agent playlist generation
+status: In Progress
+assignee: []
+created_date: '2026-04-04 08:54'
+updated_date: '2026-04-04 08:56'
+labels:
+  - agent
+  - quality
+  - bug
+dependencies: []
+references:
+  - src-tauri/src/agent.rs
+  - docs/agent.md
+priority: medium
+---
+
+## Description
+
+<!-- SECTION:DESCRIPTION:BEGIN -->
+The local LLM (qwen) hallucinated when given the prompt "instrumental tracks only" -- it returned a playlist of 22 tracks (named "Instrumental Horizons") that are overwhelmingly vocal tracks (Ellie Goulding, Grimes, Weezer, Mother Mother, Cults, etc.). The model has no way to determine whether a track is actually instrumental; it just matched Last.fm tags like "instrumental", "post-rock", "electronic" and assumed those artists' tracks are instrumental.
+
+Since the LLM cannot be trusted to make this determination from Last.fm metadata alone, add a post-generation validation step that uses the existing lyrics lookup (lrclib) to check whether each track in the generated playlist actually has lyrics. If lyrics are found, the track is not instrumental and should be filtered out.
+
+This applies specifically to prompts that request instrumental/no-vocals content, not to all playlist generation.
+
+**Observed behavior (2026-04-04):**
+- Prompt: "instrumental tracks only"
+- Result: 22 tracks, nearly all with vocals
+- Artists included: Ellie Goulding, Grimes, Weezer, Cults, MGMT, Mother Mother -- none instrumental
+<!-- SECTION:DESCRIPTION:END -->
+
+## Acceptance Criteria
+<!-- AC:BEGIN -->
+- [ ] #1 When the agent prompt implies instrumental-only tracks, each candidate track is checked for lyrics before inclusion
+- [ ] #2 Check the local SQLite db for cached lyrics first -- if the track has been played before, lyrics should already be stored
+- [ ] #3 Only call the lrclib API for tracks with no cached lyrics in the db
+- [ ] #4 Tracks with lyrics found (cached or fetched) are excluded from the final playlist
+- [ ] #5 Tracks where no lyrics exist (or lrclib returns instrumental flag) are kept
+- [ ] #6 The validation does not run for non-instrumental prompts (no performance penalty for normal playlists)
+- [ ] #7 If filtering removes too many tracks, the agent is informed and can search for more candidates
+<!-- AC:END -->

…-and-think-toggle-to-scripts-agent.py.md …-and-think-toggle-to-scripts-agent.py.mdbacklog/tasks/task-307 - Add-structured-JSONL-logging-temperature-control-and-think-toggle-to-scripts-agent.py.md renamed to backlog/completed/task-307 - Add-structured-JSONL-logging-temperature-control-and-think-toggle-to-scripts-agent.py.md backlog/tasks/task-307 - Add-structured-JSONL-logging-temperature-control-and-think-toggle-to-scripts-agent.py.md renamed to backlog/completed/task-307 - Add-structured-JSONL-logging-temperature-control-and-think-toggle-to-scripts-agent.py.md

@@ -0,0 +1,71 @@
+---
+id: TASK-311
+title: >-
+  Refactor Genius to preview/refine/approve flow instead of one-shot playlist
+  creation
+status: In Progress
+assignee: []
+created_date: '2026-04-04 08:59'
+updated_date: '2026-04-04 09:07'
+labels:
+  - genius
+  - ux
+  - refactor
+dependencies: []
+references:
+  - docs/agent.md
+  - docs/genius.md
+  - crates/mt-tauri/src/agent/mod.rs
+  - app/frontend/js/components/genius-browser.js
+priority: medium
+---
+
+## Description
+
+<!-- SECTION:DESCRIPTION:BEGIN -->
+Currently, Genius one-shots playlist creation: the user enters a prompt, the agent generates tracks, and a playlist is immediately persisted to the database. There is no opportunity to preview, refine, or reject the result before it's saved.
+
+Refactor the flow into a multi-step dialogue:
+
+1. **Generate** — User submits a prompt, agent generates a candidate track list (as today)
+2. **Preview** — Display the candidate tracks in the UI *without* persisting to the database. User can see track names, artists, and any other relevant metadata
+3. **Refine** — User can provide feedback (e.g., "remove the jazz tracks", "add more Radiohead", "make it longer") which triggers another agent turn to revise the candidate list. This can repeat multiple times
+4. **Approve** — User explicitly approves the final track list, at which point the playlist is created in the database
+
+### Backend changes (Rust)
+
+- Split `agent_generate_playlist` into two phases:
+  - **Generate/refine phase**: Returns a candidate `AgentResponse` with track list but does NOT persist to database
+  - **Approve phase**: New Tauri command (e.g., `agent_approve_playlist`) that accepts the finalized track IDs and creates the playlist
+- Support multi-turn refinement: the agent needs conversational context (prior turns) to refine intelligently. Consider whether to maintain session state in the backend or pass conversation history from the frontend
+- `shuffle_spread_artists()` should run at approve time, not generate time
+
+### Frontend changes (Alpine.js)
+
+- After generation, show a preview panel with the candidate tracks (not a toast)
+- Provide UI for:
+  - Approving the playlist (triggers persist)
+  - Entering refinement feedback (triggers another agent turn with context)
+  - Canceling/discarding the candidate
+- Track list in preview should show enough metadata for the user to evaluate quality (title, artist, album, duration)
+- Consider drag-to-reorder or remove-individual-tracks in the preview
+
+### Key files
+
+- `crates/mt-tauri/src/agent/mod.rs` — agent orchestration, `agent_generate_playlist`, `parse_agent_response`, `shuffle_spread_artists`
+- `app/frontend/js/components/genius-browser.js` — Alpine.js component, generation flow
+- `app/frontend/js/api/agent.js` — Tauri IPC bridge
+- `app/frontend/views/genius.html` — Genius view template
+<!-- SECTION:DESCRIPTION:END -->
+
+## Acceptance Criteria
+<!-- AC:BEGIN -->
+- [ ] #1 Submitting a Genius prompt returns a candidate track list without creating a playlist in the database
+- [ ] #2 Candidate tracks are displayed in a preview panel showing title, artist, album, and duration
+- [ ] #3 User can enter refinement feedback that sends another agent turn with conversational context, updating the preview
+- [ ] #4 Refinement can be repeated multiple times before approving
+- [ ] #5 User can approve the candidate, which persists the playlist to the database and emits PlaylistsUpdatedEvent
+- [ ] #6 User can discard/cancel a candidate without any database side effects
+- [ ] #7 shuffle_spread_artists runs at approve time, not generate time
+- [ ] #8 Existing one-shot behavior is fully replaced (no legacy path)
+<!-- AC:END -->