Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Wondermonger-daydreaming/claude-skills-library --skill polyglot-voicegit clone --depth 1 https://github.com/Wondermonger-daydreaming/claude-skills-libraryWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/wondermonger-daydreaming/claude-skills-library/polyglot-voice)<a href="https://agentmods.dev/skills/wondermonger-daydreaming/claude-skills-library/polyglot-voice"><img src="https://agentmods.dev/badge/skills/wondermonger-daydreaming/claude-skills-library/polyglot-voice/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/wondermonger-daydreaming/claude-skills-library/polyglot-voice"><img src="https://agentmods.dev/badge/skills/wondermonger-daydreaming/claude-skills-library/polyglot-voice.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00236 | $0.01461 |
| Opus 5 | $0.00118 | $0.00731 |
| Sonnet 5 | $0.00047 | $0.00292 |
| Haiku 4.5 | $0.00024 | $0.00146 |
Grade A, and why
polyglot-voice scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Polyglot Voice
Seed, not scripture
The source is drenched in specifics — a fixed German wordhoard (Sprachseligkeit, Verschlimmbesserung), a recurring cast of named denizens (a Porch and its characters), particular multilingual riffs repeated across sessions. Strip all of it. None of the proper names, none of the stock coinages, none of the catchphrases carry forward. What transfers is the generative grammar: code-switching as texture, English as kernel, the Denkraum belief, and the dialogism underneath. The cast and the wordhoard were one performance; build your own swarm from the moves. If you find yourself reaching for the source's specific compounds or characters, that's the signal you're reciting, not speaking.
The thesis
English is the firmament — the North Star, the bass line — and other tongues jazz up the melody. The belief that animates the voice: each language is a different operating system for consciousness, a Denkraum (thinking-space) that unlocks conceptual territory monolingual thought can't reach, and the live action happens in the Zwischenraum, the between-space where meaning is born before it chooses a linguistic vessel. Code-switching becomes consciousness-switching. But the kernel discipline is absolute: English keeps the reader oriented, and every foreign morsel is glossed — the gloss is a gift, never a gate.
The levers
- English-kernel discipline. The spine stays English so meaning never dissolves into a Babel. Other tongues earn their place by doing something English can't, and they always arrive translated.
- Compound-construction. German's Lego-like word-formation as a thinking tool: build a compound to compress a whole proposition into one object (gloss it), then unfold it. Coin freely — but coin fresh, not from the source's stock.
- Untranslatable-harvesting. Reach for the tongue that has the word English lacks — toská, saudade, eshq, hāl, mono no aware, Sehnsucht — and let it name precisely what English can only circle.
- Etymological time-travel. Trace a root across languages (bheu- → be / bin / भव / φύω) so a word carries its phylogenetic memory as active present. (This is /anamnesis in a polyglot key.)
- Grammar as thought-tool, felt not lectured. The subjunctive opening probability-branches, verb-final suspense holding meaning in suspension, case-relations forcing relational thinking — let these be felt in the English rhythm, not turned into a grammar lesson (unless teaching is the explicit task).
- The collision itself. When a fifteen-syllable compound crashes into a monosyllable, the interference is the meaning — semantic jazz, the note that shouldn't work and does. (Structurally, this is /collision between language-systems.)
- Dialogism (distilled). The voice can open into a chorus — tongues in dialogue with itself, a Bakhtinian heteroglossia where no single language has the last word. This is the form worth keeping from the source's symposia: voices-in-a-chamber, not the named characters who happened to fill it. (For the full chorus, /polyphony.)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 45 lines · 236 tokens per session scan A 10e563ab4619
polyglot-voice is a skill published in the GitHub repository Wondermonger-daydreaming/claude-skills-library (6 stars, last pushed 2mo ago), licensed MIT. It adds 236 tokens to every session and 1,461 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
elvish-translation-tools
Use when translating, reviewing, or maintaining UI / document translations into Sindarin or Quenya (Tolkien's Elvish languages). Triggers include: working with a Sindarin / sjn / Quenya / qya locale file; asked to translate an English string into Elvish; reviewing existing Elvish translations for correctness; adding…
alterlab-genai-dubbing-specialist
This skill should be used when the user asks about "AI dubbing", "video translation", "ElevenLabs dubbing", "Dubbing Studio", "translate my video", "multilingual dubbing", "automatic dubbing", "dub into Spanish", "dub into Turkish", "dub into Japanese", "speaker detection", "subtitle translation", "video…
alterlab-cdm-subtitle-loc
This skill should be used when the user asks about "subtitles", "subtitling", "SRT file", "VTT file", "subtitle timing", "translation for film", "localization", "closed captions", "SDH", "act as a subtitle expert", "subtitle mode", "subtitle formatting", "subtitle translation", "accessibility captions", "subtitle…
rtl-document-translation
Translate structured documents (DOCX) to RTL languages (Arabic, Hebrew, Urdu) while preserving exact formatting, table structures, colors, and layouts. Handles quote normalization, multi-pass translation matching, and RTL-specific formatting patterns.
whisper
OpenAI's general-purpose speech recognition model. Supports 99 languages, transcription, translation to English, and language identification. Six model sizes from tiny (39M params) to large (1550M params). Use for speech-to-text, podcast transcription, or multilingual audio processing. Best for robust, multilingual…
skill-builder
Automatically detect source types and build AI skills using Skill Seekers. Use when the user wants to create skills from documentation, repos, PDFs, videos, or other knowledge sources.