The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user input which is then used to construct shell commands.
Ingestion points: The <test-app-name> and <variant-name> arguments provided by the user (identified in SKILL.md).
Boundary markers: Absent. While the skill instructs the agent to "parse" and "clean" the input, it does not provide explicit boundary markers or instructions to ignore embedded malicious content within the strings.
Capability inventory: The skill executes several shell commands including ls, cd, git, and pnpm across multiple steps.
Sanitization: Absent. There is no explicit requirement for the agent to sanitize or escape shell metacharacters before interpolation.
[COMMAND_EXECUTION]: The skill relies on shell command execution to perform its primary functions, including repository state inspection and build processes.
Executes git status --porcelain to identify modified packages.
Dynamically constructs and executes commands like pnpm run test:run <app-name> based on user input.
Navigates directories using cd based on package names derived from git output or user input.

e2e