The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): No evidence of credential harvesting, sensitive file access, or unauthorized network activity was found. The skill operates entirely on local image data using standard libraries.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): The skill depends on standard, reputable Python packages (pillow, imageio, numpy). No suspicious remote scripts or piped bash commands are present.
[Indirect Prompt Injection] (LOW): The skill is designed to process user-provided images and instructions to create GIFs.
Ingestion points: Reads user-uploaded images via PIL.Image.open as described in SKILL.md and core/gif_builder.py.
Boundary markers: Standard image format parsing acts as a boundary; no explicit NLP boundary markers for metadata-based instructions are implemented.
Capability inventory: File write access via imageio.imwrite in core/gif_builder.py and image manipulation via Pillow.
Sanitization: Uses standard library functions for image decoding and quantization which are generally resilient to prompt-based manipulation, though usual caution with untrusted binaries applies.
[Command Execution] (SAFE): The skill does not invoke system shells or execute arbitrary commands. All logic is implemented in native Python using established APIs.
[Obfuscation] (SAFE): All source code and instructions are provided in clear text. No encoding tricks, zero-width characters, or hidden payloads were detected.

slack-gif-creator