kadirnar/voicehub/add-or-validate-speech-model

Claude Code and agent skill by kadirnar/voicehub

Trust tier 3
  • community
Source check pending

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

Add, complete, or audit a VoiceHub TTS, ASR, or VAD model integration across configuration, runtime, registry, processing, serialization, provenance, licensing, tests, optimization, and Transformers-style documentation. Use whenever model packages, checkpoints, model pages, aliases, or contribution scaffolds change.

Install

npx vskill@latest install kadirnar/voicehub/add-or-validate-speech-model

SKILL.md

Add Or Validate A Speech Model

Inventory the Integration

  1. Read .ai/AGENTS.md, .ai/GOAL.md, .ai/LOOP.md, and

docs/project/adding-a-model.md.

  1. Identify the canonical registry key, uppercase-first display name, task,

architecture, configuration, processor, runtime, outputs, checkpoint source, revision, license, model page, and optimization capabilities.

  1. Verify external facts with the official model repository, paper, checkpoint

metadata, and license. Pin immutable revisions.

Complete the Contract

  1. Keep model-specific graphs, conversion, and preprocessing in the model or

architecture package.

  1. Provide lazy registry and auto-class discovery without heavy imports.
  2. Normalize inputs, outputs, configuration, loading, saving, failure behavior,

and optional dependency errors through shared VoiceHub contracts.

  1. Preserve safe checkpoint boundaries. Prefer Safetensors; use restricted,

weights-only loading only when conversion requires another format.

  1. Declare honest capabilities. Never infer support from a provider name or

silently skip unsupported behavior.

  1. Ensure every public optimization can apply, validate, restore, serialize,

and report through the universal contract.

  1. Generate the Transformers-style model page and navigation entry from the

source of truth. Do not hand-maintain duplicated model metadata.

Verify

  1. Run focused import, construction, configuration, processor, execution,

failure, serialization, registry, model-page, and optimization tests.

  1. Verify lazy discovery in a fresh process without importing heavy backends.
  2. Run CPU-safe representative execution and deterministic round trips.
  3. Use a real released checkpoint when accessible and practical. Otherwise,

record the exact inaccessible or hardware-limited gate as unverified.

  1. Run registry-wide, documentation, packaging, and broader checks in

proportion to the changed contract.

Deliver

Keep the pull request limited to one model or one shared contribution contract. Report provenance, license, checkpoint evidence, tests, skipped gates, and all generated outputs explicitly.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

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

Source

Popularity

Repository stars
119
Repository forks
11
7-day trend score
24

Stars and forks are for the whole GitHub repository, not this skill alone. 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/kadirnar/voicehub/add-or-validate-speech-model/badge)](https://verified-skill.com/skills/kadirnar/voicehub/add-or-validate-speech-model)

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