The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions attempting to override agent behavior or bypass safety filters were found. The use of terms like 'NEVER' and 'IMPORTANT' is restricted to technical constraints and best practices for developers.
[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file paths, or unauthorized network operations were detected. API calls shown in examples (e.g., '/api/data') are illustrative and follow standard patterns.
[Obfuscation] (SAFE): No Base64, zero-width characters, homoglyphs, or other obfuscation techniques are present in the documentation.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): The documentation references industry-standard libraries (MUI, TanStack, Zod, Zustand). No remote scripts are downloaded or executed. Code splitting patterns use standard dynamic imports for local features.
[Indirect Prompt Injection] (SAFE): While the skill describes data ingestion surfaces (forms and URL parameters), it explicitly promotes the use of Zod schemas for strict validation, which acts as a primary mitigation against injection attacks in the final application.
[Dynamic Execution] (SAFE): The examples demonstrate standard React component rendering and do not use unsafe methods like 'eval()' or 'exec()'. Library loading (e.g., jspdf, xlsx) is handled through standard ES6 dynamic imports.

frontend-dev-guidelines