The Agent Skills Directory

[SAFE]: The skill's primary purpose is to identify and rewrite user-facing strings (labels, error messages, etc.) to improve UX quality. All operations are confined to reading source files and modifying string literals or localization files.
[SAFE]: While the skill scans sensitive-sounding files like package.json, build.gradle, and Podfile, it does so only to detect the development stack and locate UI string files. No sensitive data like credentials or API keys are targeted or exfiltrated.
[SAFE]: The 'Self-Healing Validation' phase includes running project-specific build commands (e.g., npm run build). This is a standard development practice to ensure that the modifications to string constants or translation files did not break the project's compilation.

design-copy