segment-image

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides image segmentation functionality using the Gemini vision model. It operates on local image files and writes output masks to the file system as described in the documentation.\n- [COMMAND_EXECUTION]: The script segment.ts runs under the bun runtime to manage command-line arguments and communicate with the Gemini service.\n- [PROMPT_INJECTION]: The skill ingests untrusted data from images and prompts in segment.ts. It lacks specific boundary markers for prompt interpolation. Its capabilities include local file writing (writeFile) and directory creation (mkdir), and it applies regex-based sanitization to output filenames. This risk surface is inherent to its primary function and is handled safely within its intended context.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 04:02 AM