The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides a 'Subprocess Output' pattern in references/core/patterns.md that demonstrates how to spawn child processes and display their output.
Ingestion points: The SubprocessViewer component accepts command and args as props, which are potential entry points for untrusted data if used dynamically.
Boundary markers: The example does not include specific delimiters or instructions to ignore embedded commands in the input data.
Capability inventory: The pattern utilizes the Node.js spawn function in references/core/patterns.md, which allows for system-level command execution.
Sanitization: The provided code snippet does not demonstrate input validation or sanitization before passing props to the spawn function.
[COMMAND_EXECUTION]: The skill documents the use of the child_process.spawn method for executing system commands as part of its 'Subprocess Output' pattern in references/core/patterns.md. While this is an intended feature for CLI development, it involves the execution of external processes.

ink