i18n-ready
Skill12 starsChanged 12 days agoMITView on GitHub
Multilingual-resilient web UI, English/Spanish first-class. Use for localization, translation, text expansion, lang attributes, logical properties, RTL readiness, locale formats, bilingual copy tells, hreflang, and language toggles.
In this repository21 other files it gives its agents.
- AGENTS.mdAGENTS.md
- llms.txtllms.txt
- a11y-passSkill
- art-directionSkill
- cognitive-a11ySkill
- color-systemSkill
- component-statesSkill
- data-vizSkill
- design-system-interviewSkill
- deslop-uiSkill
- empty-statesSkill
- form-uxSkill
- humanize-copySkill
- improve-existing-websiteSkill
- micro-motionSkill
- responsive-layoutSkill
- spacing-systemSkill
- tastecheck-passSkill
- tasterollSkill
- themingSkill
- web-typographySkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Never invent locale, idiom, translation, or formality.
- Distinctive: line 32
Test the longest approved fixture.
- Very specific: line 34
Use logical inline/block properties for RTL and Intl with real locales for formats.
- Distinctive: line 48
Keep three lanes distinct: meaning (native/approved intent, formality, idiom, plural/protected wording).
- Very specific: line 83
Stop / handoff: Stop when supported locales or source meaning are undefined.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
i18n-Ready (bilingual by design, not by patch)
Design for the longest honest string; translate meaning, not words. English/Spanish is the worked pair; the method generalizes.
Make a locale contract before translating or resizing
Record a locale contract before resizing/translating:
| Boundary | Required decision/evidence |
| locale/reader | concrete BCP 47 tag, fallback/owner, formality and regional vocabulary |
| content | whole-message keys, variables/plurals, protected legal/safety language |
| mechanics | formats and data source for date/time zone/number/currency/address/name |
| stress/direction | longest approved fixture by surface; current dir, RTL/bidi/logical-property path |
Never invent locale, idiom, translation, or formality; pending source/approval is structural readiness only.
The decision order
- Set
<html lang>and inlinelangswitches. - Test the longest approved fixture; text containers/content-sized controls avoid fixed widths. Intentional truncation exposes full text.
- Use logical inline/block properties for RTL and
Intlwith real locales for formats. - Re-express native voice/idiom/formality (including tú/usted); do not calque.
- For an in-product language choice, self-label/persist the control, update
lang, and providehreflang/per-locale metadata for indexed locale routes where applicable.
Non-negotiables
- Correct language attributes, longest-string-safe layout, and logical CSS—no text images or concatenated fragments.
- Explicit-locale
Intl; native Spanish with consistent formality, accents/ñ/¿¡, andcognitive-a11ypass per locale.
Verify meaning, layout, and mechanics separately
Keep three lanes distinct: meaning (native/approved intent, formality, idiom,
plural/protected wording); layout (longest fixtures in real views/states;
pseudo-localization finds width defects only); mechanical (Intl, time zones, currencies,
inputs, lang/dir, logical properties, toggle persistence). A date changing calendar day
is a data defect.
Repair order and evidence
Require locale contract and approved fixtures first. Repair boundaries, formats, keys, wrapping, logical CSS/bidi, and native voice; record separate three-lane checks per locale/surface with reviewer/source authority.
Self-check (per language, before shipping)
- Correct
lang, logical CSS, longest-fixture layout, and no fragmented/image text? - Explicit-locale
Intl, native/formality-safe copy, and per-language cognitive pass? - Honest persisted toggle/metadata where applicable, with separate meaning/layout/mechanical proof?
How to deliver
Deliver tag/fallback, formality source, longest fixture, three-lane evidence, bidi/format
risk, and owner. Keep locales in DESIGN-SYSTEM.md; hand off voice/type/accessibility.
Reference files
references/decision-records.md— meta-patterns + ADR rules.assets/locale-fixtures.json— EN/ES keys, formats, and longest-string fixtures.
<!-- contract:v1:start -->Contract (generated)
Canonical detail: contract.json.
- Route: A UI or copy system must survive English/Spanish expansion, locale formatting, bidi, or locale-native voice.; avoid: The request is only English copy polish or a non-locale accessibility defect.
- Exclude: Do not translate word-for-word when locale-native meaning is required. (+1 in contract.json)
- Stop / handoff: Stop when supported locales or source meaning are undefined. (+1 in contract.json); receives [humanize-copy, responsive-layout, form-ux, improve-existing-website, web-typography] -> sends [a11y-pass, tastecheck-pass]
- Output: multilingual resilience repair ledger
- Evidence:
table_with_evidencewithstatus,reason,remediation,evidence,provenance.
<!-- contract:v1:end -->Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.