Agent skill

Find Unwrapped Strings

by B0und in B0und/WikiSpeedrun

Audit a Lingui project for hardcoded user-facing strings that were never wrapped in macros.

MITAuto-check passedFrontend & Design

Install Find Unwrapped Strings

skills CLI
$ npx skills add B0und/WikiSpeedrun --skill find-unwrapped-strings -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install B0und/WikiSpeedrun find-unwrapped-strings --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/B0und/WikiSpeedrun.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/find-unwrapped-strings .claude/skills/find-unwrapped-strings && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
find-unwrapped-strings
GitHub stars
133
Token cost
~2.1k tokens
SKILL.md length
1,144 words
Files
2 (incl. references)
Skills in repo
6
Repo updated
First seen
Licence
MIT

At a glance

Audit a Lingui project for hardcoded user-facing strings that were never wrapped in macros.

  • Works in 6 steps: Ensure the guardrail → Scan → Judge each hit → …
  • Asked to find untranslated
  • SKILL.md covers When to run this, Step 1 — Ensure the guardrail, Step 2 — Scan and Step 3 — Judge each hit, plus 5 more sections
  • Calls npx and npm

What it does

Find Unwrapped Strings is an agent skill from B0und/WikiSpeedrun. Audit a Lingui project for hardcoded user-facing strings that were never wrapped in macros. Use when asked to find untranslated, unwrapped, or hardcoded strings, to check i18n coverage, to audit what an i18n setup or migration missed, or when text renders in the source language after everything was supposedly translated.

Its SKILL.md is about 2.1k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files, including reference files (for example `references/eslint-config.md`).

It sits in Frontend & Design, covering Internationalization. It works with ESLint. The repository describes itself as: Wikipedia Speedrun Game. The licence is MIT.

When your agent uses it

  • Asked to find untranslated
  • Hardcoded strings
  • Check i18n coverage
  • Audit what an i18n setup

Example prompts

  • “/find-unwrapped-strings”

Requirements

  • Node.js

Workflow steps

6 steps, taken from the step headings in SKILL.md.

  1. Ensure the guardrail
  2. Scan
  3. Judge each hit
  4. Fix in bounded batches
  5. Terminate
  6. Report

What it can do on your machine

Read from SKILL.md and the folder at commit a03d655. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npx
    • npm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npx and npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Find Unwrapped Strings loads about 2.1k tokens when it runs, and up to ~4.5k if it reads all its reference files. Until then it costs about 86 tokens; SKILL.md has 1,144 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~86
When it runs · the whole SKILL.md, loaded when a task matches
~2.1k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~4.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from B0und/WikiSpeedrun at commit a03d655, republished under its MIT licence (© B0und). 1,144 words, ~2,081 tokens.

Download SKILL.mdSave it as .claude/skills/find-unwrapped-strings/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
find-unwrapped-strings
description
Audit a Lingui project for hardcoded user-facing strings that were never wrapped in macros. Use when asked to find untranslated, unwrapped, or hardcoded strings, to check i18n coverage, to audit what an i18n setup or migration missed, or when text renders in the source language after everything was supposedly translated.

Find Unwrapped Strings

When to run this

  • After a Lingui setup or a migration onto Lingui — both wrap what they touch, and neither proves the rest of the codebase was touched.
  • When the report is "the build is green and the catalogs are full, but this screen still shows English in every locale."
  • As a periodic coverage audit on a project where i18n is already established.

The misses are systematic, not random. Strings in JSX get wrapped because they look like UI; what survives is display copy that doesn't — inside data modules (export const products = [{ name: "AeroPress Go" }]), message maps in toast and error helpers, labels in config objects.

The scan is over-inclusive by design. Precision comes from judging each hit against the skip-list, never from tuning the scanner until it goes quiet. A quiet scanner is indistinguishable from a clean codebase, and only one of those is worth having.

Judge by role, never by language. The scanner flags string literals structurally and the skip-list decides which are user-facing — what natural language a string appears to be written in is not an input to that decision, and working it out is pure overhead.

Step 1 — Ensure the guardrail

Install eslint-plugin-lingui and enable no-unlocalized-strings at warn with the tuned options from references/eslint-config.md:

bash
npm install --save-dev eslint-plugin-lingui

Take the current release — it's a lint-only dev dependency, so there's nothing a pinned range protects.

This install is permanent. The audit is a one-time sweep; the rule is what stops the next regression. An audit that finds 30 strings and leaves no guardrail behind buys a few weeks.

Enable the rule explicitly. The plugin's recommended and flat/recommended presets do not include it, so extending a preset is not coverage — the reference has the detail.

If the user declines a new dev dependency, there is a one-shot mode — write the tuned config outside the repo and point ESLint at it, leaving the project untouched:

bash
npx eslint -c /tmp/lingui-audit.config.js --no-config-lookup 'src/**/*.{ts,tsx}' --format json

Offer that only after they decline, and say what it costs: the audit works, but nothing prevents the next unwrapped string and the next audit starts from zero. Recommend the install.

Either way, this step is done once a scan runs and the rule reports.

Step 2 — Scan

bash
npx eslint 'src/**/*.{ts,tsx}' --format json > /tmp/lingui-scan.json

Filter to ruleId === "lingui/no-unlocalized-strings" and reduce to file:line plus the offending source line — the reference has the jq recipe. That list is the worklist; work it top to bottom.

Keep the rule's options in the config file and run the CLI with no rule flags. An options-bearing --rule replaces the configured options rather than merging, roughly doubling the hit list with class names and log arguments — and the longer list reads as more thorough, so its garbage gets "fixed" by wrapping identifiers. The reference has the measurements.

When a scan comes back clean on a codebase you have reason to suspect, check the glob before believing it: a files pattern of src/**/*.ts never lints .tsx, and the run still exits 0.

Step 3 — Judge each hit

Every hit gets a disposition — wrapped, or skipped with a recorded reason. No hit leaves this step without one.

The skip-list lives in the lingui-best-practices skill, section "Don't Wrap Non-UI Strings" — the authority on what stays unwrapped. Read it there before dispositioning. If it's out of reach, the role test below carries the judgment.

Dispositions turn on the string's role, not its shape. "Best seller" in a data module is a badge a shopper reads — wrap it. "aero-press-go" in the same object is a URL segment — skip it. Two questions settle the ambiguous ones:

  • Could a translator change this without breaking anything? ("DRAFT" as a status value: no.)
  • If this rendered in Japanese, would that be correct or a bug? (A sku in Japanese is a bug.)

When the line alone can't answer, read the consuming site. A string's role lives where it's used, not where it's defined.

Show full SKILL.md (507 more words)Show less

Step 4 — Fix in bounded batches

Wrap per the macro decision tree in lingui-best-practices. The cases this audit turns up most:

  • JSX content → Trans
  • A string inside a component (attribute, alert, function argument) → useLingui() + t`…`
  • A string outside any component — data module, module-level map, config object → msg descriptor from @lingui/core/macro, resolved at the consuming site.

That last case is what this audit exists for, and it's the one that gets fixed wrongly. t at module scope resolves once, at import time, against whatever locale happened to be active, and then never changes. Define with msg, resolve where it renders:

ts
// src/data/products.ts — definition
import { msg } from "@lingui/core/macro";

export const products = [
  { sku: "sku-1001", slug: "aero-press-go", name: msg`AeroPress Go` },
];
tsx
// consuming component — resolution
const { t } = useLingui();
<h2>{t(products[0].name)}</h2>;

The field's type changes from string to MessageDescriptor, so TypeScript points at every consuming site — follow it. In non-React code with no render moment, i18n._(descriptor) resolves at call time, which is what a toast helper wants.

Add translator comments as you wrap, per enhanced-message-context: data-module labels and one-word badges are its must-comment cases, arriving with no surrounding UI to infer meaning from.

Then, once per round:

bash
npx lingui extract --clean
npx lingui compile

and re-scan from Step 2.

Bounds: at most 2 fix rounds, or ~40 files. On hitting either, stop and report. A human can rerun the audit; a human cannot review a 200-file diff.

Step 5 — Terminate

Stop when one of these is true, and say which:

  1. The scan is clean.
  2. A round wrapped zero new strings — every remaining hit is a confirmed false positive. Grow the rule's configured ignores from those specific cases, per the reference's discipline, so the guardrail stays quiet while still reporting everything else. Severity stays at warn, the rule stays enabled, and hits get ignores rather than eslint-disable comments.
  3. The bound is exhausted (2 rounds / ~40 files).

Step 6 — Report

Every hit from every scan lands in one of three buckets, stated with file:line:

  • Wrapped — what changed, and which macro.
  • Skipped — with the skip-list reason (slug, enum value, log output).
  • Residual — still open, reviewed or not, listed individually so the next run starts here.

Report the guardrail state too: severity, and which ignores you added with the evidence that earned each.

Account for every hit before claiming the audit is finished. "I wrapped 16 strings" is not a result; "wrapped 16, skipped 16 as identifiers and log output, 0 residual, rule left at warn" is. Unreported residuals are the one outcome this skill exists to prevent — the misses were invisible in the first place precisely because nobody wrote them down.

Invoked by another workflow?

An orchestrator (a setup skill, a TMS-connection skill) may supply scope (src/features/checkout only, or the files a migration touched), extra skip-list entries from a project rules file (brand names, product codes), or known misses ("an earlier pass missed the data layer").

Honor all three. The pipeline and the reporting contract don't change: scope narrows what you scan, never what you report. Findings outside the given scope are reported as out-of-scope, not omitted.

  • lingui-best-practices — owns the skip-list and the macro decision tree that Steps 3 and 4 run on.
  • enhanced-message-context — translator comments for the short, context-free labels this audit surfaces.

© B0und, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (references) in .agents/skills/find-unwrapped-strings of B0und/WikiSpeedrun.

  • SKILL.md
  • references/eslint-config.md

Open the folder on GitHubat commit a03d655

Compare with similar skills

Find Unwrapped Strings next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Find Unwrapped Strings compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Find Unwrapped Strings this skillB0und/WikiSpeedrun133—~2.1kAutomated safety check: PassMIT
Impeccablebestofjs/bestofjs3.1k27 repos~2.6kAutomated safety check: PassMIT
Chatbox i18n Translatorchatboxai/chatbox42k—~508Automated safety check: PassGPL-3.0
Internationalization Workflow with i18niOfficeAI/AionUi33k1 repos~1.9kAutomated safety check: PassApache-2.0
Enforce Rules For I18nmoeru-ai/airi50k—~1.5kAutomated safety check: PassMIT
Claude Desktop Chinese Localizationjavaht/claude-desktop-zh-cn7.5k—~1.6kAutomated safety check: PassMIT

Similar skills

  • Impeccable

    bestofjs/bestofjs

    A skill your agent uses when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a…

    3.1k GitHub starsUsed in 27 repos~2.6k tokens
    Frontend & DesignAuto-check passed
  • Chatbox i18n Translator

    chatboxai/chatbox

    Translates new or changed i18n keys from a Chatbox Pro diff, staged changes or a commit range, writing the locale JSON files directly with a built-in glossary.

    42k GitHub stars~508 tokensUpdated 14 days ago
    Frontend & DesignAuto-check passed
  • Standards for keeping all user-facing text translatable: read the i18n config first, use namespaced keys, reuse shared strings and follow the key naming rules.

    33k GitHub starsUsed in 1 repo~1.9k tokens
    Frontend & DesignAuto-check passed
  • Review pending AIRI translations on Crowdin in a batch, then sync them into the repository.

    50k GitHub stars~1.5k tokensUpdated today
    Frontend & DesignAuto-check passed
  • Claude Desktop Chinese Localization

    javaht/claude-desktop-zh-cn

    Adds missing Simplified and Traditional Chinese translations to the Claude Desktop Chinese patch across three layers, then checks how many mappings actually hit.

    7.5k GitHub stars~1.6k tokensUpdated 4 days ago
    Frontend & DesignAuto-check passed
  • Taro UI Guide

    jd-opensource/taro-ui

    Guides installing, configuring, styling and using taro-ui (At* components) in Taro apps for WeChat, Alipay, H5 and React Native.

    4.7k GitHub stars~1.3k tokensUpdated 14 days ago
    Frontend & DesignAuto-check passed

More from B0und/WikiSpeedrun

  • Lingui Best Practices

    B0und/WikiSpeedrun

    Implement internationalization with Lingui in React and JavaScript applications.

    133 GitHub starsUsed in 2 repos~4.2k tokens
    Auto-check passed
  • Lingui Framework Setup

    B0und/WikiSpeedrun

    Set up Lingui in a React framework. An agent skill from B0und/WikiSpeedrun.

    133 GitHub stars~1.9k tokensUpdated 19 days ago
    Auto-check passed
  • Migrate I18next To Lingui

    B0und/WikiSpeedrun

    Migrate i18next/react-i18next projects to Lingui. An agent skill from B0und/WikiSpeedrun.

    133 GitHub stars~3.1k tokensUpdated 19 days ago
    Auto-check passed
  • Enhanced Message Context

    B0und/WikiSpeedrun

    Add translator comments to Lingui messages so translations are accurate.

    133 GitHub stars~3.3k tokensUpdated 19 days ago
    Auto-check passed
  • Swc Plugin Compatibility

    B0und/WikiSpeedrun

    Diagnose and fix Lingui SWC plugin compatibility errors with Next.js, Vite, Rspack, or other SWC runtimes.

    133 GitHub stars~2.3k tokensUpdated 19 days ago
    Auto-check passed

Works with

Questions about Find Unwrapped Strings

What does Find Unwrapped Strings do?

Audit a Lingui project for hardcoded user-facing strings that were never wrapped in macros. Find Unwrapped Strings is an agent skill from B0und/WikiSpeedrun. Audit a Lingui project for hardcoded user-facing strings that were never wrapped in macros.

When should I use Find Unwrapped Strings?

Find Unwrapped Strings fits situations like: asked to find untranslated; hardcoded strings; check i18n coverage; audit what an i18n setup.

How do I install Find Unwrapped Strings in Claude Code?

Run `npx skills add B0und/WikiSpeedrun --skill find-unwrapped-strings -a claude-code`. Or copy the skill folder (.agents/skills/find-unwrapped-strings in B0und/WikiSpeedrun) into .claude/skills/find-unwrapped-strings in your project. Claude Code loads it when a task matches its description.

How do I install Find Unwrapped Strings in Codex?

Run `npx skills add B0und/WikiSpeedrun --skill find-unwrapped-strings -a codex`. Or copy the skill folder (.agents/skills/find-unwrapped-strings in B0und/WikiSpeedrun) into .agents/skills/find-unwrapped-strings in your project. Codex loads it when a task matches its description.

Can I use Find Unwrapped Strings in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add B0und/WikiSpeedrun --skill find-unwrapped-strings -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/find-unwrapped-strings, .gemini/skills/find-unwrapped-strings, .github/skills/find-unwrapped-strings and .opencode/skills/find-unwrapped-strings in your project.

What does Find Unwrapped Strings need to run?

Going by SKILL.md and its folder, Find Unwrapped Strings needs the command-line tools its instructions call (npx and npm). Our summary lists: Node.js.

Does Find Unwrapped Strings access the network?

SKILL.md contains no URLs. Its commands use npx and npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Find Unwrapped Strings safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Find Unwrapped Strings use?

Find Unwrapped Strings is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Find Unwrapped Strings use?

About 2.1k tokens (SKILL.md is roughly 8.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.4k tokens, read only when the agent opens those files.

What are the alternatives to Find Unwrapped Strings?

Skills that share tags, products or a category with Find Unwrapped Strings: Impeccable (bestofjs/bestofjs, 3.1k stars), Chatbox i18n Translator (chatboxai/chatbox, 42k stars), Internationalization Workflow with i18n (iOfficeAI/AionUi, 33k stars) and Enforce Rules For I18n (moeru-ai/airi, 50k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Find Unwrapped Strings?

B0und (a GitHub user) maintains it in B0und/WikiSpeedrun, which has 133 GitHub stars. The repository holds 6 skills in this directory. The repository was last updated on September 19, 2026.

Source: B0und/WikiSpeedrun on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.