The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses user-supplied text to build and execute shell commands via the Nx CLI. This creates a vulnerability where a malicious user could attempt to inject command arguments.
Ingestion points: User instructions for project scaffolding defined in the skill's purpose.
Boundary markers: None identified; the skill lacks instructions for the agent to sanitize or validate input project names or options.
Capability inventory: Execution of commands like nx generate, nx lint, nx test, and nx build which can modify the filesystem and run code.
Sanitization: Absent; the skill directs the agent to map user intent directly to shell arguments.
[COMMAND_EXECUTION]: The skill's core functionality is built around executing shell commands for repository management, including nx list, nx generate, nx format, and various verification tasks.
[DYNAMIC_EXECUTION]: The skill uses a dynamic Node.js evaluation to programmatically find plugin files.
Evidence: The use of node -e "console.log(require.resolve(...))" to locate generator JSON files.

nx-generate