The Agent Skills Directory

[PROMPT_INJECTION]: The skill documents an attack surface for indirect prompt injection by providing templates for Kafka consumers that process external data. This is an inherent part of the skill's purpose to facilitate messaging and is addressed via best practice recommendations.
Ingestion points: The generated {YourModel}Processor class in SKILL.md ingests raw external data from $message->payload.
Boundary markers: The templates do not use specific boundary markers for the data, as the generated code is intended for structured application logic rather than direct re-insertion into an LLM context.
Capability inventory: The generated code across all templates is limited to standard Laravel model operations and does not include high-risk capabilities like shell execution, dynamic code evaluation, or arbitrary file access.
Sanitization: The skill explicitly includes a 'Consumer Best Practices' section in SKILL.md that instructs developers to always validate message structures and implement error handling.

ensi-kafka