The Agent Skills Directory

[PROMPT_INJECTION]: No bypass or override instructions detected. The instructions follow natural instructional language for a technical template.
[DATA_EXPOSURE_AND_EXFILTRATION]: No sensitive file access or credential harvesting was identified. The usage of the 'NEXT_LOCALE' cookie follows security best practices by defaulting to 'httpOnly: true' to prevent client-side script access.
[OBFUSCATION]: The source code and documentation do not contain any Base64, hex-encoded strings, zero-width characters, or homoglyph substitutions.
[REMOTE_CODE_EXECUTION]: No remote scripts are downloaded or executed. The package installation processes use standard registries for well-known dependencies.
[PRIVILEGE_ESCALATION]: The skill does not attempt to use 'sudo' or modify system-level permissions. Build scripts are restricted to standard project directory cleanup.
[PERSISTENCE_MECHANISMS]: No persistence logic (such as shell profile modification or cron jobs) is present.
[METADATA_POISONING]: Metadata fields are consistent with the skill's stated purpose and functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes 'Accept-Language' headers and cookies. It mitigates injection risks by validating all input against a hardcoded 'supportedLanguages' whitelist using canonicalization logic in 'utils.ts'.
[TIME_DELAYED_ATTACKS]: No logic was found that gates execution based on dates, times, or external environmental triggers.
[DYNAMIC_EXECUTION]: The skill uses static build-time compilation with 'tsup' and does not utilize 'eval()', 'exec()', or other dynamic code execution functions.

nextjs-locale-monorepo