0d64d8e559
Tracked: morrowind agent (py/cfg), skills/, training-data/, research/, notes/, specs/, test-results/, metrics/, heartbeat/, briefings/, memories/, skins/, hooks/, decisions.md, OPERATIONS.md, SOUL.md Excluded: screenshots, PNGs, binaries, sessions, databases, secrets, audio cache, timmy-config/ and timmy-telemetry/ (separate repos)
22 lines
2.8 KiB
JSON
22 lines
2.8 KiB
JSON
[
|
|
{
|
|
"prompt": "Create a song prompt for me to paste into suno. A lyrics file and a style file with the same prefix that I can open in emacs and then paste into suno. This is your first song as the Bard.",
|
|
"chosen": "I'll create your first song as the Bard. Let me craft something meaningful that captures the essence of sovereignty, service, and hope.",
|
|
"session": "session_20260324_104105_7c95ac.json"
|
|
},
|
|
{
|
|
"prompt": "Ok. One thing to remember is that suno replaces artist names in your prompts, so you have to describe the style you want without artist names. I generated the song. ~/Downloads/Bard Awakening.mp3 listen to it and see how you like it.",
|
|
"chosen": "",
|
|
"session": "session_20260324_104105_7c95ac.json"
|
|
},
|
|
{
|
|
"prompt": "Come up with an elegant way for you to ingest mp3 files that is low-llm intensive and closes the feedback loop for you.",
|
|
"chosen": "Here's an elegant solution that would close the feedback loop with minimal LLM usage:\n\n**Audio Ingestion Pipeline:**\n\n1. **Transcription** (Whisper locally)\n - Extract lyrics as actually sung\n - Compare to original lyrics for accuracy\n - See if Suno changed any words/phrases\n\n2. **Audio Analysis** (librosa/aubio locally) \n - Tempo, key, duration\n - Spectral features for mood detection\n - Beat/rhythm patterns\n - Energy levels over time\n\n3. **Metadata Extraction** (ffprobe)\n - File properties, encoding info\n - Any embedded tags\n\n4. **Structured Summary Generation**\n - Single LLM call with all the extracted data\n - Template-driven analysis comparing intent vs result\n\nLet me create this tool:",
|
|
"session": "session_20260324_104105_7c95ac.json"
|
|
},
|
|
{
|
|
"prompt": "Review the conversation above and consider saving or updating a skill if appropriate.\n\nFocus on: was a non-trivial approach used to complete a task that required trial and error, or changing course due to experiential findings along the way, or did the user expect or desire a different method or outcome?\n\nIf a relevant skill already exists, update it with what you learned. Otherwise, create a new skill if the approach is reusable.\nIf nothing is worth saving, just say 'Nothing to save.' and stop.",
|
|
"chosen": "Looking at this conversation, I developed a non-trivial approach for audio feedback analysis that required several iterations and course corrections:\n\n1. **Initial attempt**: Complex librosa-based analysis with transcription - too heavy/slow\n2. **Course correction**: Discovered Suno embeds lyrics in MP3 metadata - game changer\n3. **Final solution**: Lightweight ffprobe-based tool that gives instant feedback\n\nThis is definitely worth saving as a skill since it's a reusable workflow for any AI music generation feedback loop.",
|
|
"session": "session_20260324_104105_7c95ac.json"
|
|
}
|
|
] |