The Agent Skills Directory

[COMMAND_EXECUTION]: The test script tests/test_image_gen.py uses subprocess.run to call image generation scripts within the skill's directory to verify their behavior.\n- [EXTERNAL_DOWNLOADS]: The scripts/generate_fal.py script downloads generated image content from remote URLs returned by the fal.ai API.\n- [PROMPT_INJECTION]: The skill processes user-supplied prompts without sanitization, creating a potential surface for indirect prompt injection.\n
Ingestion points: The --prompt CLI argument in scripts located in the scripts/ directory.\n
Boundary markers: No delimiters or instruction-level isolation are provided when sending prompts to the remote models.\n
Capability inventory: The skill possesses network access and file write permissions to save generated images.\n
Sanitization: User prompts are passed directly to external APIs without filtering or validation.\n- [SAFE]: Authentication is managed through standard environment variables (OPENAI_API_KEY, GOOGLE_API_KEY, FAL_KEY), avoiding hardcoded secrets.\n- [SAFE]: The skill interacts with well-known and reputable AI services using official or widely-accepted client libraries.

image-gen