The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): The skill instructions define a professional persona ('senior frontend engineer') to guide architectural decisions. It contains no attempts to bypass safety filters or ignore prior instructions.
[CREDENTIALS_UNSAFE] (SAFE): No hardcoded API keys, tokens, or secrets were found. Authentication patterns are handled via a local useAuth hook as per best practices.
[EXTERNAL_DOWNLOADS] (SAFE): The skill references standard, reputable industry libraries such as React, MUI, TanStack Query, and Zod. No untrusted or remote script execution (e.g., curl | bash) is present.
[DATA_EXFILTRATION] (SAFE): No code patterns or network operations were found that would transmit sensitive local data to unauthorized external domains.
[COMMAND_EXECUTION] (SAFE): The skill is strictly limited to frontend React development and does not invoke system-level commands, shell scripts, or subprocesses.
[INDIRECT_PROMPT_INJECTION] (SAFE): The skill demonstrates secure data ingestion patterns by explicitly recommending Zod for schema validation and input sanitization in forms.
[DYNAMIC_EXECUTION] (SAFE): While the skill utilizes React.lazy and dynamic import() for code splitting, these are standard performance optimizations using static paths and do not involve executing untrusted dynamic strings.

frontend-dev-guidelines