audio-transcriber
Transform audio recordings into professional Markdown documentation with intelligent summaries using LLM integration
50
Quality
30%
Does it follow best practices?
Impact
85%
2.17xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/antigravity-audio-transcriber/SKILL.mdEvaluation results
86%
56%Podcast Transcription Automation Script
Faster-Whisper model parameters and output file naming
Faster-Whisper import
0%
100%
Whisper fallback
0%
100%
CPU device param
0%
100%
int8 compute type
0%
100%
VAD filter enabled
0%
100%
Word timestamps enabled
0%
100%
Default model is base
100%
100%
Timestamp output naming
0%
0%
Segment timestamps in output
40%
80%
No hardcoded API keys
100%
100%
Language auto-detect
100%
100%
Without context: $0.3964 · 6m 38s · 23 turns · 70 in / 5,180 out tokens
With context: $1.1191 · 6m 50s · 35 turns · 7,047 in / 10,895 out tokens
86%
49%Meeting Documentation from Raw Transcript Data
Markdown report template structure
Top-level title
0%
100%
Metadata section heading
50%
100%
Metadata as table
100%
100%
All metadata fields present
100%
100%
Meeting minutes heading
100%
100%
Participants subsection
25%
100%
Topics subsection
25%
100%
Decisions with checkmark prefix
0%
0%
Action items with checkboxes
0%
100%
Segment timestamps MM:SS format
0%
71%
No large prose dump
100%
100%
Footer attribution
0%
100%
Without context: $0.1831 · 2m 12s · 11 turns · 15 in / 2,778 out tokens
With context: $0.5972 · 4m 29s · 18 turns · 7,030 in / 6,577 out tokens
85%
33%Universal Audio Meeting Summarizer Tool
Auto-detection, LLM CLI integration, and RISEN prompt
Claude CLI first
100%
100%
GH Copilot fallback
100%
100%
Claude invoked via stdin
0%
100%
RISEN prompt structure
0%
41%
RISEN documents in design notes
0%
0%
Rich library for UI
100%
100%
Temp file cleanup
100%
100%
Faster-whisper preferred
0%
100%
Detection order documented
100%
100%
No hardcoded API keys
100%
100%
Timestamp file naming
0%
100%
Without context: $0.5421 · 6m 50s · 24 turns · 31 in / 9,163 out tokens
With context: $0.7600 · 5m 10s · 23 turns · 3,196 in / 9,308 out tokens
- Repository
- boisenoise/skills-collections
- Commit
5c5ae21
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.