What problem does it solve? ImageMagick's behavior varies by build, and silent failures (wrong output format, missing coders, destructive in-place overwrites) are common. This Skill lets an agent turn plain-English image requests into correct, verified magick commands while detecting missing delegates, policy restrictions, and silent format fallbacks before they corrupt results. ## Core Features & Use Cases - Full image operation coverage: inspection and EXIF, format conversion and compression, resizing, cropping, rotation, color and tone, transparency, compositing, watermarking, text and drawing, effects, montages, animation, batch processing, image comparison, and PDF/SVG/PSD rasterization. - Runtime capability detection: preflight checks of -list format, delegates, and policy, plus round-trip probes, so the agent never promises formats the local build cannot actually write (e.g., a broken HEIC coder that silently writes a PNG). - Safety guardrails: protects originals from destructive mogrify runs, treats untrusted files as an attack surface with FORMAT: prefixes and resource limits, and validates every output's real format, dimensions, and alpha state. - Use Case: Ask "Get this JPEG under 200 kilobytes" or "Remove the GPS location from these photos" and receive a verified command whose output is checked with identify and semantic assertions. ## Quick Start Ask the agent to resize every JPEG in a folder to 1200 pixels wide into a new output directory using ImageMagick.