The Agent Skills Directory

[SAFE]: No security issues detected. The skill uses standard shell commands (mkdir, touch, ls) and curl for local development tasks which are appropriate for its stated purpose.
[COMMAND_EXECUTION]: Shell commands used are limited to local file system management and communication with a local development server (localhost:3000), presenting no significant risk in this context.
[PROMPT_INJECTION]: The skill ingests content from a local URL for testing, which is a necessary functional requirement. This surface is considered safe as it operates within a local development environment. Ingestion points: test-content-path parameter. Boundary markers: Not required for local dev content. Capability inventory: Local file writing and shell execution. Sanitization: Handled by the agent's underlying safety layers during local task execution.

building-blocks