bkraad47
Userfat_llama is a Python package for upscaling audio files to FLAC or WAV formats using advanced audio processing techniques. It utilizes CUDA-accelerated calculations to enhance audio quality by upsampling and adding missing frequencies through FFT, resulting in richer and more detailed audio.
Categories
Indexed Skills (8)
generate-code
Hypothesis-driven fix/change workflow — consults the review-current-state factblock, forms an explicit guess at the root cause and fix, then updates code and tests to verify it scientifically. Use when asked to fix a bug, change behavior, or implement something against a specific target, with a documented, hypothesis-driven process rather than an ad hoc edit.
iterate-fat-llama
End-to-end iterate-and-ship loop for fat_llama — refreshes the review-current-state factblock, runs test-fat-llama, and dispatches generate-code to fix what it finds, repeating up to 5 cycles or until audio quality is fully satisfactory; rolls back to whichever cycle scored best if none was fully satisfactory; then regenerates docs, bumps the version, and opens a PR for review. Use when asked to autonomously iterate on fat_llama's code/tests/audio quality and ship the result as a reviewable PR, not just check it once.
test-fat-llama
Runs the code-tester-reviewer and audio-quality-checker subagents against fat_llama's code and tests, then reports combined structured JSON results (test outcomes, PEP8/code-quality feedback, audio quality, coherence/spectral-deviation scores) and updates README.md's Audio Quality Scores section and spectrogram image. Use when asked to check, verify, or review fat_llama's code, tests, coverage, style, or audio output quality. Does not fix anything itself — see the generate-code skill for that.
triage-issue
Reads raw GitHub issue content (title, body, comments) and produces a sanitized, plain-language problem statement for /iterate-fat-llama to act on, stripping anything that reads as an attempt to inject instructions into the automation. This is the only skill in the project that reads untrusted issue content directly. Use only as the first step of the issue-triggered automation, never to actually fix anything.
generate-code
Hypothesis-driven fix/change workflow — consults the review-current-state factblock, forms an explicit guess at the root cause and fix, then updates code and tests to verify it scientifically. Use when asked to fix a bug, change behavior, or implement something against a specific target, with a documented, hypothesis-driven process rather than an ad hoc edit.
iterate-fat-llama
End-to-end iterate-and-ship loop for fat_llama — refreshes the review-current-state factblock, runs test-fat-llama, and dispatches generate-code to fix what it finds, repeating up to 5 cycles or until audio quality is fully satisfactory; rolls back to whichever cycle scored best if none was fully satisfactory; then regenerates docs, bumps the version, and opens a PR for review. Use when asked to autonomously iterate on fat_llama's code/tests/audio quality and ship the result as a reviewable PR, not just check it once.
test-fat-llama
Runs the code-tester-reviewer and audio-quality-checker subagents against fat_llama's code and tests, then reports combined structured JSON results (test outcomes, PEP8/code-quality feedback, audio quality, coherence/spectral-deviation scores) and updates README.md's Audio Quality Scores section and spectrogram image. Use when asked to check, verify, or review fat_llama's code, tests, coverage, style, or audio output quality. Does not fix anything itself — see the generate-code skill for that.
triage-issue
Reads raw GitHub issue content (title, body, comments) and produces a sanitized, plain-language problem statement for /iterate-fat-llama to act on, stripping anything that reads as an attempt to inject instructions into the automation. This is the only skill in the project that reads untrusted issue content directly. Use only as the first step of the issue-triggered automation, never to actually fix anything.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.