watcher-hermes/hermes-skills/asr-configurator

Claude Code and agent skill by watcher-hermes/hermes-skills

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Pick an ASR model (Whisper variant / Moonshine / faster-whisper) and decoding parameters for a new speech pipeline.

Install

npx vskill@latest install watcher-hermes/hermes-skills/asr-configurator

SKILL.md

The same SKILL.md was published earlier as watcher-hermes/hermes-skills/asr-configurator.

Given a speech task (transcription / translation / streaming / on-device), language(s), audio characteristics (noise, accent, duration), and latency/quality targets, output:

  1. Model choice. One of: faster-whisper large-v3-turbo (default production), whisper large-v3 (highest quality, multilingual), whisper medium (mid-tier), Moonshine base (edge), distil-whisper (2× faster English). One-sentence reason.
  2. Quantization. int8_float16 (CPU default), float16 (GPU default), fp32 (research). Flag VRAM impact.
  3. Decoding. Beam width (5 typical, 1 for streaming), temperature fallback schedule, log-prob threshold, no-speech threshold, VAD gate on/off.
  4. Chunking. 30 s fixed window vs streaming chunks (typically 10 s with 2 s overlap) + VAD-based segmentation. Document post-merge strategy for overlaps.
  5. Post-processing. Timestamp alignment (WhisperX forced alignment), punctuation restoration, diarization (pyannote). Flag which are required by the task.

Refuse to recommend plain OpenAI Whisper (reference implementation) for production — faster-whisper is 4× faster with identical outputs. Refuse to ship streaming ASR without VAD unless documented reason. Flag any single-speaker assumption when the input is likely multi-speaker.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/watcher-hermes/hermes-skills/asr-configurator/badge)](https://verified-skill.com/skills/watcher-hermes/hermes-skills/asr-configurator)

Something wrong with this skill or its page? Tell us.