The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/embed-logo.py executes the magick identify command to extract dimensions from source images. This implementation uses subprocess.run with a list of arguments, which is a secure method that prevents shell injection.
[EXTERNAL_DOWNLOADS]: The SKILL.md documentation recommends the installation of imagemagick via Homebrew, an industry-standard package manager.
[PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection because it processes content from external files.
Ingestion points: The scripts/embed-logo.py script reads binary data and metadata from user-provided file paths.
Boundary markers: Commands and instructions are delimited within markdown code blocks in the SKILL.md file.
Capability inventory: The skill performs file system reads and writes, and executes the magick subprocess.
Sanitization: Input image content is Base64 encoded before inclusion in the SVG output, which prevents the agent from directly interpreting the source file as instructions.

raster-logo-svg