The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill's implementation guide in SKILL.md utilizes the new Function() constructor to dynamically compile and execute string-based expressions. This pattern is used for both syntax validation and testing, effectively enabling runtime execution of arbitrary JavaScript code provided as input data.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and evaluate n8n expressions that may originate from untrusted external workflow data or user inputs, creating a surface for indirect injection.
Ingestion points: Found in SKILL.md within the validateExpressionSyntax and testExpression functions, which take an expression string as input.
Boundary markers: There are no explicit delimiters or boundary instructions provided to the execution environment to isolate the expression content from the underlying system logic.
Capability inventory: The skill leverages new Function to execute the input strings, providing a powerful capability for code execution.
Sanitization: The provided code snippets perform basic removal of template markers ({{ }}) but lack robust sandboxing or deep sanitization of the JavaScript content before execution.

n8n-expression-testing