From image-annotation
Batch-convert images to WebP for web-optimised output. Originals are left untouched; converted files land in <dir>/webp/<stem>.webp by default. Supports lossy quality control or lossless mode.
How this skill is triggered — by the user, by Claude, or both
Slash command
/image-annotation:batch-to-webpThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Convert one or many images to WebP via `cwebp`. Useful after annotation when the user is preparing screenshots for a blog post, docs site, or anywhere bandwidth matters.
Convert one or many images to WebP via cwebp. Useful after annotation when the user is preparing screenshots for a blog post, docs site, or anywhere bandwidth matters.
Never overwrite or delete the source files. Output defaults to <dirname>/webp/<stem>.webp next to each input — a single sibling subfolder collects the converted set, leaving originals clean.
Convert specific files:
bash "$CLAUDE_PLUGIN_ROOT/scripts/to-webp.sh" \
[--quality 85] [--lossless] [--out-dir DIR] \
FILE [FILE ...]
Or convert by glob:
bash "$CLAUDE_PLUGIN_ROOT/scripts/to-webp.sh" \
--glob "/path/to/dir/*.png" [--quality 85] [--out-dir DIR]
--quality 85 — solid balance of size vs visual fidelity for screenshots.--lossless for diagrams, UI screenshots with sharp text, or anything where compression artefacts on edges would be visible.--out-dir overrides the default webp/ sibling and dumps everything into one folder (handy for blog uploads).The script prints each converted file path on stdout and a converted=N skipped=M summary on stderr. Surface both to the user.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub danielrosehill/claude-code-plugins --plugin image-annotation