The Agent Skills Directory

[SAFE]: No malicious patterns, prompt injections, or unauthorized data access detected. The skill's behavior is consistent with its stated purpose of assisting in scientific figure preparation.
[COMMAND_EXECUTION]: The skill contains Python code snippets designed to perform local file operations, specifically reading image files via the Pillow library. These operations are intended for validating image metadata (resolution, mode, format) and do not involve shell injection or execution of untrusted commands.
[EXTERNAL_DOWNLOADS]: The skill references official documentation from the American Association for Cancer Research (aacrjournals.org). These are informational links to a well-known academic service and do not involve remote code execution or automated downloads.
[INDIRECT_PROMPT_INJECTION]: The skill defines an interface for processing external data in the form of image files.
Ingestion points: The image_path argument in the check_cancer_res_resolution and validate_cancer_res_figure functions (SKILL.md).
Boundary markers: None identified.
Capability inventory: File reading for image metadata extraction (PIL.Image.open). No network operations, file writing, or privilege escalation capabilities are present.
Sanitization: Standard image decoding is handled by the Pillow library. Given the lack of dangerous downstream capabilities, this surface is considered safe.

cancer-research-figure-guide