The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user-provided images, creating an attack surface for indirect prompt injection where malicious instructions could be embedded in image metadata or pixel data.
Ingestion points: The skill uses PIL.Image.open() to process user-uploaded files as documented in SKILL.md and implemented in various templates.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between image data and potentially malicious embedded text instructions.
Capability inventory: The skill has the capability to write files to the local disk via the GIFBuilder.save() method and performs complex image transformations.
Sanitization: The provided scripts do not perform any sanitization or filtering of image metadata or content to prevent the processing of embedded instructions.
[COMMAND_EXECUTION]: The SKILL.md documentation includes instructions for the user to execute shell commands using the pip package manager to install necessary dependencies.

slack-gif-creator