The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses grep via shell to scan the project source code for hardcoded strings and specific UI attributes.
[COMMAND_EXECUTION]: Phase 2 utilizes node -e to run inline JavaScript for comparing content across translation files.
[PROMPT_INJECTION]: The skill processes untrusted codebase data (source files, translation files) and has access to shell and file-write tools. Evidence Chain: 1. Ingestion points: codebase source files and translation files; 2. Boundary markers: Absent; 3. Capability inventory: Shell access (grep), dynamic JS execution (node -e), and file system write access; 4. Sanitization: Absent.

audit-i18n