Localize Orchestrator Skill
Use this skill when the user asks for end-to-end localization of newly added or changed UI strings.
Purpose
Run the localization pipeline in a fixed order so missed strings are found, wrapped, translated, and embedded.
Required Order
- •
localize-search- •Find potentially unlocalized user-facing strings.
- •Produce a candidate list with file path, line reference, detected string, and review status.
- •If no candidates remain after filtering, stop and report a no-op result.
- •
localize-update- •Apply
loc("...")updates for the selected candidate strings. - •Keep this step source-edit only (no translation or embedding).
- •Apply
- •
localize-strings- •Run extraction + translation workflow to update
localize/strings.tsv. - •Embed localized resources and validate generated outputs.
- •Run extraction + translation workflow to update
Execution Rules
- •Do not skip steps unless the user explicitly asks.
- •Preserve
localize-searchexclusions and review behavior (including brand/product-name review before exclusion). - •Keep edits minimal and scoped to localization.
- •Use existing project scripts and avoid new dependencies.
- •Avoid duplicate command execution across steps.
Reporting Requirements
Return a concise combined report with:
- •Search findings summary (including excluded/reviewed cases).
- •Files changed by update step.
- •Translation/localization artifacts changed by strings step.
- •Commands run and any failures/follow-ups.
Skill References
- •Search:
.github/skills/localize-search/SKILL.md - •Update:
.github/skills/localize-update/SKILL.md - •Strings:
.github/skills/localize-strings/SKILL.md