The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to inspect local project files for design context. These actions are limited to informational gathering and do not involve sensitive system access or network calls.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by reading external data from project files. 1. Ingestion points: Data is read from globals.css, tailwind.config.*, and DESIGN-BRIEF.md using find and grep. 2. Boundary markers: None identified; no instructions exist to isolate file data or warn the agent about untrusted content. 3. Capability inventory: Includes shell command execution and local file writing to the project root. 4. Sanitization: No sanitization or filtering of the file contents is performed before the agent processes the information.

design-grill