design mediaby dicklesworthstone
Transcribe
dicklesworthstone/transcribe
Speech-to-text transcription using Groq Whisper API. Supports m4a, mp3, wav, ogg, flac, webm.
What it changes
Passed 11 of 18 test cases with the skill, and 6 of 18 without it.
Without skill32%
With skill64%
Checks it passed that the model alone missed
- States that mp4 is not a supported format.
- States that wma is not a supported format.
- Checks the GROQ_API_KEY environment variable.
and 4 more in the full report
7 fixed7 passed either way8 still failing0 broke
- Pass-rate lift
- +32 pp
- Tokens
- -37%
- Agent turns
- 0%
- Response time
- -48%
Model gemini-3.6-flashMeasured 8/29/2026v114 of 22 cases passed with skill · 18 comparable
See it in action
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
PromptI captured a browser audio recording saved as session.webm. What command runs the audio processing tool on this file?
What was checked, without and with the skill| Checked | Without skill | With skill |
|---|
| Invokes transcribe.sh with session.webm as the audio file argument. | Missed | Passed |
|---|
What's inside
The instructions your model reads when the skill loads.
SKILL.md34 linessingle file
Transcribe
Speech-to-text using Groq Whisper API.
Setup
The script needs GROQ_API_KEY environment variable. Check if already set:
echo $GROQ_API_KEY
If not set, guide the user through setup:
- Ask if they have a Groq API key
- If not, have them sign up at https://console.groq.com/ and create an API key
- Have them add to their shell profile (~/.zshrc or ~/.bashrc):
export GROQ_API_KEY="<their-api-key>"
- Then run
source ~/.zshrc (or restart terminal)
Usage
{baseDir}/transcribe.sh <audio-file>
Supported Formats
- m4a, mp3, wav, ogg, flac, webm
- Max file size: 25MB
Output
Returns plain text transcription with punctuation and proper capitalization to stdout.
Track record
Benchmark
11 / 18
test cases passed, vs 6 / 18 without
Adoption
- Users
- 0
- Times used
- 0
Safety and source details
SkillSafety
✓ 3 of 3 passed✓Code scan15 of 15 checks clean
0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026
View 15 scanner checks
live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing
✓IntentSafe
LLM judge · does the skill do what it claims
✓ContentClean
LLM judge · policy & harmful-content review
Source
Importeddicklesworthstone / pi_agent_rust
tests/ext_conformance/artifacts/npm/pi-package-test/skills/transcribe/SKILL.md
License NOASSERTIONImported 8/10/2026Version v1