The Agent Skills Directory

[SAFE]: The skill provides documentation for performing pixel-level image comparisons using standard Python libraries. No malicious code, obfuscation, or exfiltration patterns were identified.- [INDIRECT_PROMPT_INJECTION]: The skill describes processing external image files. While this presents a theoretical attack surface for indirect injection via image data, the documented tool performs localized pixel comparisons and outputs structured results, limiting the potential for malicious instruction execution. * Ingestion points: Image files provided as command-line arguments. * Boundary markers: None. * Capability inventory: Local file writing. * Sanitization: Not specified in the documentation.

pixel-diff