How to use Gemini to localize a technical blog
Connie Leung used a Gemini skill to translate technical Markdown with Taiwan glossaries, protected code and URLs, then saved a translation with link checks.
dev.to
Original prompt
Shared by the source author. See source notes for context.
use the technical-localization-expert skill to translate @mastering_live_sports_data_with_gemini_3.md to Traditional Chinese and write the translation to chinese-blogs/translated3.md
Reach people exploring AI agents
See placements →What happened
A Traditional Chinese Markdown translation was saved with an audit and reported matching, reachable links, ready for human review.
Reported by Connie Leung on DEV CommunityHow the agent did it
- Leung installed regional terminology references and a link-validation script as a skill.
- Gemini activated the skill and translated a named Markdown source into the requested file.
- It checked protected assets and links, then returned a saved translation and audit report.
Where a person stepped in
Leung created terminology references, activated the requested skill and retained the translated Markdown for review and publication.
Connie Leung used a custom Gemini CLI skill to translate an English technical post into Traditional Chinese for Taiwanese developer communities. The workflow combined regional terminology references with an executable link check, addressing more than the prose translation alone.
After restarting the CLI with the skill installed, Leung requested translation of a named Markdown post into chinese-blogs/translated3.md. Gemini asked to activate the skill, loaded the Taiwan terminology files and read the source post into context. The references distinguished local technology vocabulary from alternatives used in other Chinese-speaking regions.
The skill protected code, variable names, backtick blocks and URLs while translating link labels and image descriptions. It then ran a Node.js validator to compare source and target assets and check links. The demonstrated run reported matching, reachable links and produced an audit summary alongside the saved translation.
Leung retained the file for further review and publication. The source therefore establishes a completed localization artifact with deterministic asset checks, not independent proof that every translation choice was linguistically perfect. Its recovery policy also separated translation-created link errors from problems already present in the English source, limiting correction attempts instead of looping indefinitely.
Prompt
Original prompt
Source: DEV Community · Connie Leung
Frequently asked questions
How were regional terms chosen?
The skill loaded Taiwan-specific technology glossaries.
Were code and URLs translated?
No. They were protected while prose, link text and image descriptions were localized.
What was produced?
A Traditional Chinese Markdown file and an action audit with reported successful link checks.
More details
About this account
After restarting the CLI with the skill installed, Leung requested translation of a named Markdown post into chinese-blogs/translated3.md. Gemini asked to activate the skill, loaded the Taiwan terminology files and read the source post into context. The references distinguished local technology vocabulary from alternatives used in other Chinese-speaking regions.
First-person syndicated article and API opened; linked personal canonical page was inaccessible. This verified author-published copy describes an executed translation and saved file. Original prompt is verbatim. Language quality is reported, not independently reviewed. Original prompt excerpt retained from the source.
Credit: Connie Leung on DEV Community
Source published 19 Feb 2026
Source checked 14 Oct 2026
Tools and topics
Gemini CLI · Node.js · Markdown
Technical localization · Link validation
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.