The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill implements an Azure Functions runtime that processes external HTTP event data. This creates a surface for indirect prompt injection where malicious instructions could be embedded in request payloads processed by the agent.
Ingestion points: The event and request objects passed to the run() methods in Function.java and BoxLang handler files (Lambda.bx, Products.bx, etc.).
Boundary markers: No specific boundary markers or "ignore instructions" delimiters are enforced in the provided boilerplate code.
Capability inventory: The runtime has capabilities for file system access, response serialization, and environment variable access as part of the BoxLang execution environment.
Sanitization: No explicit sanitization or filtering of input data is implemented in the provided starter templates.
[COMMAND_EXECUTION]: The documentation includes instructions for executing shell commands and build scripts to manage the application lifecycle.
Evidence: Usage of ./gradlew test, ./gradlew azureFunctionsRun, ./gradlew azureFunctionsDeploy, and Azure CLI commands like az login and export for environment variables.
[EXTERNAL_DOWNLOADS]: The skill references and downloads project scaffolding from external sources.
Evidence: The instruction to git clone https://github.com/ortus-boxlang/boxlang-starter-azure-functions.git fetches the starter template from the vendor's repository.

boxlang-runtime-azure-functions