i18n
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use find, grep, and ripgrep to analyze the project's translation usage and structure.\n- [EXTERNAL_DOWNLOADS]: The documentation references well-known Ruby gems including rails-i18n, i18n-tasks, mobility, and route_translator for localization tasks.\n- [DYNAMIC_EXECUTION]: The references/testing.md file includes a test example using YAML.load_file to verify the validity of locale files, which is a standard practice for validating local configuration.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for handling user-controlled locale inputs.\n
- Ingestion points: Untrusted data is ingested from params[:locale] and HTTP_ACCEPT_LANGUAGE in SKILL.md and references/locale-switching.md.\n
- Boundary markers: The instructions focus on application-level logic separation rather than LLM-specific delimiters.\n
- Capability inventory: The skill uses Bash for Rails commands and Write/Edit for file system modifications.\n
- Sanitization: The skill explicitly recommends sanitizing inputs by checking them against an allowlist of available_locales to prevent path traversal.
Audit Metadata