claude-gif-edit

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute system utilities such as gifsicle, ffmpeg, and ffprobe for frame-level operations and video filtering. It also invokes local scripts and a Python environment located within ~/.claude/skills/claude-gif/ and ~/.video-skill/ to perform specialized GIF transformations.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external media files (INPUT.gif) and accepts user-defined parameters for text overlays, cropping, and resizing.
  • Ingestion points: INPUT.gif file content and user-supplied strings for the drawtext filter.
  • Boundary markers: Absent in the command templates.
  • Capability inventory: Extensive use of shell execution via the Bash tool to perform file writes and media processing.
  • Sanitization: The skill demonstrates safe practices by parsing dimensions using grep and sed from gifsicle --info output rather than relying on unvalidated user input for coordinate math, although raw user text interpolation in ffmpeg filters represents a potential injection surface if not handled carefully by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 07:47 AM