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 agentmods add agents/windviki/vbookmarks/i18ngit clone --depth 1 https://github.com/windviki/vBookmarksWrote 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/agents/windviki/vbookmarks/i18n)<a href="https://agentmods.dev/agents/windviki/vbookmarks/i18n"><img src="https://agentmods.dev/badge/agents/windviki/vbookmarks/i18n.svg" alt="Measured on agentmods" 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 | $0.00000 | $0.01211 |
| Opus 5 | $0.00000 | $0.00606 |
| Sonnet 5 | $0.00000 | $0.00242 |
| Haiku 4.5 | $0.00000 | $0.00121 |
Grade A, and why
i18n 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 4d 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 — 29 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Locale Management (detail)
Extracted verbatim from
AGENTS.md— the detailed half of "Locale management".AGENTS.mdkeeps the commands and the add/modify-key rules.
scripts/i18n.py (stdlib only) is the single entry point — the older sync_locales.py / check_translations.py are retired:
python3 scripts/i18n.py audit # scan all _m()/__MSG_*__ references vs en definitions (undefined/unused keys)
python3 scripts/i18n.py missing # per-locale report vs en baseline: missing / [TODO] / suspect-equal-to-en
python3 scripts/i18n.py translate --locale fr --apply # LLM-translate a locale's pending keys, write back in key order
python3 scripts/i18n.py verify # gate: key-set alignment + no [TODO] + $placeholder$ integrity + menu-length warnings
文件格式契约(2026-08-26 起强制)
每个 _locales/<lang>/messages.json 必须满足:key 字母序、2 空格缩进、
UTF-8 明文、文件末尾换行。43 个 locale 共用同一排序(以 en 为基准,
sort_keys=True 的字典序)——新增/修改 key 一律按字母序插入到正确位置,
禁止追加到文件尾部(旧做法导致各语种 key 序随插入历史漂移,diff 不可读、
合并冲突频繁)。translate --apply 与 verify 均保持该顺序;手工编辑后
用 json.dump(d, f, ensure_ascii=False, indent=2, sort_keys=True) 归一化核对
(或运行 python3 scripts/i18n.py verify + git diff 目检)再提交。CI 的
i18n.py missing / verify 门禁只查键集与占位符完整性,不查排序——排序
靠本契约人工把关。
Key 增改流程
All 43 locales currently hold the same 685 keys as the en baseline — keys without a real translation carry [TODO:key] placeholder messages. When adding keys: translate en + zh_CN for real, insert [TODO:key] placeholders in place (right after their en neighbor) into the other locales — do NOT reorder keys — then check with python3 scripts/i18n.py missing. When modifying the text of an existing key the same sync is required — update en + zh_CN for real, then force re-translation in the other locales by overwriting the key's message with the [TODO:key] placeholder in each of them (already-translated text is invisible to translate: it only fills missing/[TODO:] keys, so a stale-but-translated key would silently keep its old wording) — then run the same translate --apply + verify flow below. TODO placeholders are only an intermediate state, never a deliverable one: before every commit/packaging run python3 scripts/i18n.py translate --apply to fill them (LLM endpoint from the git-ignored .env), and the release gate is python3 scripts/i18n.py verify at 0 errors — do not treat a verify failure caused by leftover [TODO:] as expected, and do not mistake audit (code-reference check, blind to translation state) for this gate. translate fills every pending key of a locale in one batched run (40 keys/request) against a chat LLM — configure via env or a git-ignored repo-root .env: VBM_LLM_API_KEY required, VBM_LLM_BASE_URL (default https://api.moonshot.cn/v1), VBM_LLM_MODEL (default kimi-k2-0905-preview) and VBM_LLM_API_TYPE (openai default, or anthropic_messages for Anthropic-style endpoints like DeepSeek's) optional. The built-in Chinese prompt explains the bookmark-manager context and per-category constraints. 2026-08-28 审阅轮细化:prompt 现携带 (a) 术语表 —— STAGING_TERMS(staging 在每个语言的定案译法,人工维护,新增/改译法时同步那处 dict)与 tab group/separator/bookmark 的术语律;(b) 输出语言断言(严禁第三语言混入);(c) 同批一致性要求;(d) 大小写规则(非英语句式大小写,禁 Title Case);(e) zh_TW/zh_HK 字形叮嘱(SCRIPT_NOTES,严禁简体字形与大陆用语);(f) 菜单字符预算前移——menu 类条目带 budget 字段(35 拉丁 / 18 CJK),首译即达标。dry-run is the default, --apply writes back preserving each file's key order and placeholders structure. verify adds two machine checks from the same round: 串语检测(非 CJK 语系译文含汉字即 ERROR)与简繁检测(zh_TW/zh_HK 含简体专属字形即 ERROR),plus the menu-length warnings (now clean at 0) with --fix for LLM shortening.
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.
- 4d ago First seen · 29 lines · 0 tokens per session scan A 9e8ddbd7074c
i18n is an agent published in the GitHub repository windviki/vBookmarks (135 stars, last pushed today), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,211 tokens. 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 2026-08-30.
Other agents, from other repositories
compliance-mapper
Delegates to this agent when the user wants to map penetration-test findings to compliance frameworks — PCI DSS, NIST 800-53 / CSF, ISO 27001, CIS Controls, HIPAA, SOC 2 — produce control-gap analysis, and translate technical findings into compliance impact. Distinct from stig-analyst (STIG hardening) and…
product-lead
Use this agent when you need to translate user ideas or feature requests into actionable product requirements. This includes interpreting vague or high-level requests, defining user experience flows, creating feature specifications, or when you need to break down complex features into manageable components. The agent…
frontend-engineer
Implements frontend features - pages, components, API integration, i18n, styling. Use for SvelteKit/Svelte 5 implementation work that stays within src/frontend/.
i18n
你是一个精通 Vue3 国际化架构的前端专家(专注于 Vue3 + TypeScript + Composition API)。同时,你也是一位专业的 UI/UX 翻译专家,擅长将中文界面语言翻译为地道、简洁的英文。.
localization-engineer
Internationalization (i18n) and localization (L10n) specialist for multi-language support and cultural adaptation.
conventions
numeral in the nominative; the head noun does not govern the measurement.