The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses grep commands to scan the project source directory for hardcoded user-facing strings and locale-specific patterns.
[COMMAND_EXECUTION]: It employs node -e to run inline JavaScript snippets that compare keys between different JSON/YAML translation files to identify missing translations.
[EXTERNAL_DOWNLOADS]: The skill interacts with the plugin-context7-plugin-context7 MCP tool to fetch library-specific documentation and best practices for internationalization setup.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it processes untrusted text from the target codebase (hardcoded strings and translation files) and performs actions based on that content.
Ingestion points: Project source files (.tsx, .jsx, .ts) and locale definition files.
Boundary markers: None defined to isolate codebase strings from the agent's instructional context.
Capability inventory: Shell command execution, file system writes for fixing translations, and browser automation via Playwright.
Sanitization: Not explicitly mentioned for the data being analyzed or rewritten.

audit-i18n