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.
git clone --depth 1 https://github.com/roshaw/claude-translation-agencyWrote 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/commands/roshaw/claude-translation-agency/translate-audit)<a href="https://agentmods.dev/commands/roshaw/claude-translation-agency/translate-audit"><img src="https://agentmods.dev/badge/commands/roshaw/claude-translation-agency/translate-audit/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/commands/roshaw/claude-translation-agency/translate-audit"><img src="https://agentmods.dev/badge/commands/roshaw/claude-translation-agency/translate-audit.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 | $0.00399 | $0.04366 |
| Opus 5.5 | $0.00160 | $0.01746 |
| Sonnet 5.5 | $0.00080 | $0.00873 |
| Haiku 4.5 | $0.00040 | $0.00437 |
Grade A, and why
translate-audit 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 — 238 lines — stays where its author put it; the contents beside it link to each section on GitHub.
translate-audit
A read-only audit for a multi-language project, in two modes:
- Coverage (default) — is anything missing? A cheap, deterministic set comparison across every language. No agents, no LLM cost. Answers, per language: what is missing, where it is present (which other languages carry it, and the value — so you can see what it should be), and how to fix it.
- Quality (
--deep) — is what's there any good? Spawns thetranslate-leadto run its C1–C7 review over the existing translations and report polish opportunities (weak terminology, machine-translation stiffness, register/parity drift) plus correctness bugs, each with a severity and a suggested rewrite. This mode costs tokens (it reads and LLM-reviews content), so it is scoped/ sampled by default — see "Deep mode" below.
It never edits a file itself — both modes only read and report (deep mode's --fix hands the
flagged files to /translate, which does the writing). Fixing is always /translate's job; this skill
tells you exactly what to run.
Why a separate skill from
/translate?/translatehas a completeness gate, but it only checks the languages in its run scope and only against the source. This audit is broader and symmetric: it compares every language against the union of all languages, so it also catches keys that exist in Bulgarian or German but are missing from the source (English) — a class/translatestructurally can't surface, because it never adds keys to or edits the source.
What this skill does NOT do
- Translate or edit anything. No file is written except the report(s). To close gaps or apply
polish, run
/translate(this skill prints the exact command;--deep --fixinvokes it for you). - Add keys to the source, or "fix" a source gap. It flags keys missing from the source (with the value other languages use) and leaves the decision to you — add it to the source, or delete the orphan from the other languages.
- Judge quality in the default (coverage) mode. Coverage checks presence/completeness only.
Quality judgment happens only in
--deep, and even then the skill reports; it does not rewrite. - Commit or push.
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 · 238 lines · 0 tokens per session scan A a8e74fe02c6a
translate-audit is a command published in the GitHub repository roshaw/claude-translation-agency (2 stars, last pushed 1mo ago), licensed MIT. It adds 399 tokens to every session and 4,366 once invoked, about $0.0016 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 2026-09-16.
Other commands, from other repositories
start
Run a parallel persona-swarm brainstorm on a topic. Spawns 4-8 persona-distinct subagents in parallel, synthesizes their outputs into consensus / dissent / open questions / recommended next move.
audit
Audit closed tickets for regressions against today's code (unchanged, evolved, implemented wrong, not implemented), plus duplicate and groupable open ones.
pr
Take the current branch from uncommitted work to an open pull request: commit outstanding changes with the git-commit rules if asked, push with upstream tracking, write the PR title and body with the pr-description skill from the full branch diff, link the issue, and open it with gh against the right base. Confirms…
plan-localization
Plan localisation covering extraction, translation workflow, and the layout and format issues that follow.
support-model
Write a support model as Markdown in the user's repository. Produce designed HTML or PDF only when asked.
design-doc
Write a design doc / RFC as Markdown in the user's repository. Use when invoked as /document-design-system:design-doc. Produce designed HTML or PDF only when asked.