The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external image files and their associated metadata, which represents a potential surface for indirect instructions to enter the agent's context.\n
Ingestion points: The skill ingests data via the input_dir parameter in scripts/batch_convert.py and the input file path in scripts/optimize_web.py.\n
Boundary markers: The instructions do not include specific delimiters or directives to the agent to ignore or isolate instructions that might be embedded in image metadata (like EXIF tags).\n
Capability inventory: The skill possesses file system write capabilities through Image.save() and directory creation via Path.mkdir().\n
Sanitization: While the skill provides functions to strip metadata, it does not perform validation or sanitization of image content to prevent potential injection attacks targeting the agent itself.\n- [EXTERNAL_DOWNLOADS]: The skill references the installation of standard image processing dependencies.\n
Evidence: Documentation suggests installing well-known and trusted libraries including pillow, pillow-heif, and pillow-avif-plugin from public package registries.

image-converter