The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains a directive in SKILL.md instructing the agent to "fix the scripts directly" if they fail. This encourages autonomous self-modification of the Python scripts in the scripts/ directory by the AI agent during runtime.
[CREDENTIALS_UNSAFE]: The load_env function in scripts/generate.py is configured to search for and read .env files from several locations outside the skill's own directory, including the user's home directory (~/.claude/skills/.env and ~/.claude/.env), which are sensitive locations typically used for storing platform and API credentials.
[PROMPT_INJECTION]: The scripts/generate.py script interpolates user-supplied data into an AI generation prompt template (LOGO_PROMPT_TEMPLATE) without sufficient sanitization or the use of protective boundary markers. This creates an attack surface for indirect prompt injection.
Ingestion points: args.prompt, args.brand, and args.brand_context parameters in scripts/generate.py.
Boundary markers: Absent; user input is directly concatenated into the prompt string.
Capability inventory: The script has the ability to write files to the local disk and make authenticated network requests to the Google Gemini API.
Sanitization: No input validation or escaping is performed on the user-provided strings.
[EXTERNAL_DOWNLOADS]: The skill requests the installation of the google-genai Python package, which is the official SDK provided by Google for interacting with Gemini models.

logo-design