Vibe is a desktop application and command-line tool that transcribes audio and video locally on a user's device, turning speech into text and subtitle or document files. People use it for private transcription, translation, summaries, speaker labeling, and batch processing across supported desktop systems and models. The catalogue skill and instruction support working with Vibe.
Borrowing it
Nothing to install: this file belongs to thewh1teagle/vibe. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/thewh1teagle/vibe/main/.claude/skills/translate/SKILL.mdgit clone --depth 1 https://github.com/thewh1teagle/vibeWrote 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/thewh1teagle/vibe/translate)<a href="https://agentmods.dev/skills/thewh1teagle/vibe/translate"><img src="https://agentmods.dev/badge/skills/thewh1teagle/vibe/translate/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/thewh1teagle/vibe/translate"><img src="https://agentmods.dev/badge/skills/thewh1teagle/vibe/translate.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 | $0.00069 | $0.01253 |
| Opus 5.5 | $0.00028 | $0.00501 |
| Sonnet 5.5 | $0.00014 | $0.00251 |
| Haiku 4.5 | $0.00007 | $0.00125 |
Grade A, and why
translate 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 1mo 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 — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Translate
i18n/desktop/<locale>.json app strings 22 locales (desktop: true)
i18n/website/<locale>.json site strings 15 locales (website: true)
i18n/changelog/<version>.md release notes English at the top level,
i18n/changelog/<locale>/<version>.md translations under a locale dir
i18n/docs/en-US/<slug>.md docs pages same shape
i18n/docs/<locale>/<slug>.md
i18n/locales.json is the roster — filter on desktop / website. Never edit
en-US; it is the source. A missing translation falls back to English whole, so
partial coverage is safe to ship.
1. Find what is missing
uv run scripts/check_i18n.py # catalogs, changelog and docs coverage
For catalogs, the per-locale gap:
uv run python - <<'PY'
import json, pathlib
en = json.load(open("i18n/desktop/en-US.json"))
for p in sorted(pathlib.Path("i18n/desktop").glob("*.json")):
if p.stem == "en-US": continue
missing = [k for k in en if k not in json.load(open(p))]
if missing: print(p.stem, len(missing))
PY
2. Fan out — one subagent per locale
Write the brief once to a scratch file, then spawn one cheap subagent
(model: sonnet) per locale in a single message, each pointing at that file
and naming only its own locale and output path. Never translate locale after
locale yourself; that is hours of serial work for no gain.
The brief must carry these rules — every one of them was a real defect at some point:
- Placeholders byte-identical:
{count},{version},{name}… Never translate what is inside braces, never reorder them. - Structural markers stay English: the
## New/## Improved/## Fixedheadings are markers the page maps to translated badges. Same headings, same order, same bullet count as the source. - Frontmatter:
versionanddatebyte-identical to the English file; onlytitleis translated. Never write asource:line — step 4 stamps it. - UI paths are translated, including inside backticks.
Settings → Modelsnames menu items in an app that is translated, so it becomesהגדרות → מודלים— backticks and the→kept. Look the words up ini18n/desktop/<locale>.jsonso the docs match the app. - Everything else in backticks stays English, byte for byte: filenames,
flags, env vars, quoted error text, and placeholders inside example URLs —
vibe://download?url=<any model url>is a string the reader copies, not prose. - Real backticks, not
「」or""— the page renders backticks as code. - Verbatim: emoji,
(#1234)refs,@handles, URLs and link targets (translate link text only), version numbers, and product/technical names (Vibe, Whisper, ggml, FFmpeg, AVX2, CoreML, Vulkan, macOS…). Never drop a contributor credit. - Match the existing locale file for tone and terminology — tell the agent to read two or three finished files in its own locale first.
- Keep UI labels short; keep terse source terse.
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.
- 1mo ago First seen · 108 lines · 69 tokens per session scan A c74a85e754ca
translate is a skill published in the GitHub repository thewh1teagle/vibe (8,009 stars, last pushed 8d ago), licensed MIT. It adds 69 tokens to every session and 1,253 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-29.
translate compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| translatethewh1teagle/vibe | Skill | 8.0k | 8d ago | MIT | A | 69 | Claude Code |
| tauri-expertpersonamanagmentlayer/pcl | Skill | 44 | 11d ago | Apache-2.0 | A | 61 | Claude Code |
| desktop-applicationsbobmatnyc/claude-mpm-skills | Skill | 78 | 3mo ago | MIT | A | 19 | Claude Code |
| rust-desktop-applicationsbobmatnyc/claude-mpm | Skill | 158 | 2mo ago | none | A | 21 | Claude Code |
| freyamarc2332/freya | Skill | 3.3k | yesterday | MIT | A | 35 | Claude Code, Codex |
| ax-rust-signatureax-llm/ax | Skill | 3.0k | 3d ago | Apache-2.0 | A | 36 | Claude Code, Codex |
| maintaining-modding-environmentshashgraph-online/awesome-codex-plugins | Skill | 1.5k | yesterday | Apache-2.0 | A | 65 | Claude Code, Codex |