The Agent Skills Directory

[DYNAMIC_EXECUTION]: The scripts/render.mjs script uses the node:vm module to compile bundled scripts and generated HTML content to verify syntax correctness during the bundling process. This involves compiling source code strings at runtime, which is a form of dynamic execution.\n- [INDIRECT_PROMPT_INJECTION]: The skill transforms untrusted input from questions.json into formatted text summaries intended for the agent's context. This creates a surface for indirect prompt injection if the input data contains malicious instructions.\n
Ingestion points: Data from questions.json is read by scripts/render.mjs and embedded into the HTML structure.\n
Boundary markers: The generated text output does not use specific delimiters or instructions to separate user-provided content from agent instructions.\n
Capability inventory: The script performs local file system operations including creating, writing, renaming, and linking files to manage output and locks.\n
Sanitization: The skill employs safeJSON (escaping <) and htmlText encoding to sanitize data before embedding it into the HTML structure, reducing the risk of XSS within the generated file.

grilling-viz