The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines several ways to execute shell commands using pnpm for testing purposes. These include running all tests, specific test files, and watch mode within the apps/core directory. It also demonstrates how to set environment variables like NODE_OPTIONS for memory management.
[INDIRECT_PROMPT_INJECTION]: The skill utilizes an $ARGUMENTS placeholder to pass user input directly into shell commands (e.g., pnpm test -- $ARGUMENTS). This identifies a potential attack surface where untrusted data could be used to inject additional shell flags or sub-commands. However, the skill's configuration limits tool access to specific pnpm test patterns, which acts as a mitigating control.

run-test