The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides implementation examples for Retrieval-Augmented Generation (RAG) that are susceptible to indirect prompt injection. Untrusted data retrieved from a database is directly interpolated into LLM prompts without sanitization.
Ingestion points: Data is fetched from the documents table via vector similarity search in the RagService class within SKILL.md.
Boundary markers: The implementation uses a simple --- delimiter to separate chunks in the $context variable, which provides minimal protection against adversarial content.
Capability inventory: The skill demonstrates defining agents (e.g., DataAnalyst) that are granted access to tools like SearchProducts and GetSalesStats in SKILL.md.
Sanitization: There is no evidence of content filtering, escaping, or explicit 'ignore embedded instructions' directives applied to the retrieved context before it is passed to the LLM.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install an external package using composer require laravel/ai from the official Laravel vendor repository.

laravel-ai