slack-gif-creator
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of user-uploaded images via the
PIL.Image.openmethod. \n - Ingestion points: User-provided images are loaded and processed as frames as described in SKILL.md. \n
- Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions potentially embedded in image content or metadata. \n
- Capability inventory: The skill performs image transformations and writes files to the local system in core/gif_builder.py. It does not possess network access or arbitrary command execution capabilities. \n
- Sanitization: There is no explicit sanitization or filtering of image metadata (EXIF/XMP) before processing.- [METADATA_POISONING]: The LICENSE.txt file contains a copyright notice for 'Anthropic, PBC', which differs from the skill author 'davila7' listed in the skill metadata. This is likely a result of utilizing a standard license template without updating the copyright header and is considered a best-practice violation rather than a security threat.- [SAFE]: The core logic implemented in core/gif_builder.py, core/frame_composer.py, and core/easing.py uses standard mathematical functions and industry-standard libraries including Pillow and imageio for animation. No obfuscation, unauthorized privilege escalation, or persistence mechanisms were detected.
Audit Metadata