oma-voice

Local-first text-to-speech and speech-to-text via the Voicebox MCP server. Generates speech from cloned or preset voice profiles for agent notifications, content voiceovers, and audio asset creation, and transcribes audio files for meeting notes or memos. Runs entirely on-device with no cloud, no API keys, no per-call cost. Use for voice generation, TTS, STT, transcription, voiceover, narration, dictation, audio asset work.

Quality

98%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Quality

Content

96%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is lean, highly actionable, and well-sequenced with validation checkpoints and recovery flows. Its only weakness is progressive disclosure: the References section names files that are not actually shipped in the bundle, so a downstream reader hits dead links.

Suggestions

Ship the referenced bundle files (resources/voice-matrix.md, prompt-tips.md, execution-protocol.md, checklist.md, and config/voice-config.yaml) or remove their links so navigation is end-to-end valid.

Add a one-line note next to each reference indicating what it covers (e.g., "engine matrix: pick Kokoro/Piper by language and footprint") so a reader can decide whether to load it without opening the file.

Treat "verified against Voicebox 0.5.0" as a versioned anchor and add a short note on what to re-verify when Voicebox's MCP tool surface changes, so the version pin does not silently go stale.

Dimension	Reasoning	Score
Conciseness	The body is dense and table-driven, assumes Claude knows MCP/HTTP/audio concepts, and avoids explaining what TTS or a voice profile is; minor restatements between the canonical command path and the per-tool notes do not rise to over-explanation.	5 / 5
Actionability	Fully executable guidance with real tool names and parameter shapes ("MCP voicebox_speak { text, profile, language?, engine?, personality? }"), exact endpoints ("GET http://127.0.0.1:17493/audio/{generation_id}"), a concrete MCP registration command, and copy-paste invocation examples covering the common TTS/STT/notification cases.	5 / 5
Workflow Clarity	A clear PREPARE→ACQUIRE→ACT→VERIFY→FINALIZE sequence carries explicit validation (VERIFY scene, manifest-field checks, 5000-char/30-min checkpoints, exit codes) and a feedback loop (MCP tool-name drift → re-run tools/list and update the cache); the destructive/batch cap does not apply since outputs write to fresh timestamped directories.	5 / 5
Progressive Disclosure	A dedicated, well-signaled, one-level-deep References section (resources/voice-matrix.md, prompt-tips.md, execution-protocol.md, checklist.md, config/voice-config.yaml) gives good structure with no inlined bulk or nested chains, but the referenced files are absent from the bundle on disk — a minor navigation gap that keeps it off the top anchor.	4 / 5
	Total	19 / 20 Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is exemplary: it covers what the skill does and when to use it with concrete trigger terms and synonyms, and it stakes out a distinct local-only niche. No vague fluff, over-claims, or voice/person issues are present.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions across both directions of the skill — "Generates speech from cloned or preset voice profiles", "transcribes audio files for meeting notes or memos", "agent notifications, content voiceovers, and audio asset creation" — giving comprehensive coverage analogous to the rubric's PDF extract/fill/merge/convert example.	5 / 5
Completeness	Explicitly answers both what ("Local-first text-to-speech and speech-to-text via the Voicebox MCP server. Generates speech... transcribes audio files...") and when ("Use for voice generation, TTS, STT, transcription, voiceover, narration, dictation, audio asset work."), satisfying the explicit-trigger-clause requirement that otherwise caps completeness at 3.	5 / 5
Trigger Term Quality	Comprehensive natural-term coverage with synonyms and shorthand: "text-to-speech", "speech-to-text", "TTS", "STT", "transcription", "voiceover", "narration", "dictation", "voice generation" — matching the rubric's synonym-and-extension anchor.	5 / 5
Distinctiveness Conflict Risk	A clear local-only niche is carved by "Local-first", "via the Voicebox MCP server", and "Runs entirely on-device with no cloud, no API keys" — actively disambiguating from cloud TTS skills and minimizing wrong-skin triggering.	5 / 5
	Total	20 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: first-fluke/oh-my-agent
Path: .agents/skills/oma-voice/SKILL.md
Commit: 032c988

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

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.