The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied requirements to generate functional frontend and backend code, creating an attack surface where maliciously crafted requirements could attempt to influence the logic of the implemented features.
Ingestion points: User-provided feature requirements and technical design specifications ingested during triggers like 'implement feature' or 'build feature' as defined in SKILL.md.
Boundary markers: Absent; the skill does not specify explicit delimiters or 'ignore' instructions for user-provided data within the primary workflow.
Capability inventory: The skill possesses the capability to generate code for database operations, API endpoints, and network-integrated frontend components across multiple files.
Sanitization: The skill strongly mitigates this risk by mandating 'MUST DO' security constraints, including the use of Pydantic/Zod schemas for input validation, output encoding for XSS prevention, and parameterized queries to prevent SQL injection at every layer of the application.
[EXTERNAL_DOWNLOADS]: The skill's reference materials for deployment and containerization include references to external resources from well-known technology services.
Evidence: The Dockerfile template in references/backend-patterns.md utilizes the node:18-alpine official image, and the CI/CD pipeline in references/integration-patterns.md references official GitHub Actions (actions/checkout, actions/setup-node). These resources are standard, official tools from recognized providers.

fullstack-guardian