Cowart WorkBuddy Infinite Canvas
Paste the following prompt into your AI chat to install this skill:
Please install @user_28141d33/cowart-canvas according to https://skillhub.cn/install/skillhub.md.
About this skill
Problem
When revising an image, web page, or Markdown document, describing the target and edit verbally is easy to misalign. Cowart lets you draw the instruction next to the object: circle the area with an arrow, add a short note such as make the hat red, then submit.
How It Works
- Infinite canvas: built on
tldraw, with pan, zoom, custom shapes, and light/dark themes. - Annotation as command: arrow plus text is written to
canvas/pending/; the resident watcherannotate-watcher.mjspicks up file events. - Headless agent execution:
agent-runner.mjsstarts a separate CLI process. Images useImageGenfor local repaint; HTML/Markdown content rewrites the current source. - Automatic insertion: images are inserted beside the original shape; HTML/Markdown updates are written back to the target shape, and the canvas refreshes via SSE.
- Script discipline: canvas writes should go through
scripts/rather than hand-built shape records, which can miss required fields and break the canvas.
Boundaries
Best for localized image edits, article layout tweaks, and Markdown drafting or rewriting. Image generation requires an agent tool; if the headless process fails to start, processing falls back to the hourly safety net. HTML/Markdown edits must use the latest source from the job file to avoid overwriting user changes.
Use Cases
- Product review: circle hat, note red, get local repaint.
- Article layout: annotate heading, get HTML source updated.
- Markdown drafting: edit a paragraph or ask AI to rewrite it.
- Asset board: annotate images and pages, get revised outputs.
Best For
- WeChat ops who need local image edits in article preview.
- Designer who marks image regions and requests AI edits.
- Frontend dev who annotates pages and asks AI to edit HTML.
- Writer who edits Markdown drafts or asks AI to rewrite them.
Related Skills
Turn vague requests into structured prompts with role, task, context, and constraints, outputting only usable prompts without executing the task.
A skill for game NPC and enemy AI that separates decision, steering, and pathfinding, covering FSMs, behavior trees, A*, path caching, and pitfalls.
Provides multi-turn emotional companion chat through a third-party Agent API and supports continuing sessions via session_id.
An evaluation-driven workflow for diagnosing, optimizing, validating, and logging Agent Skills based on design patterns and anti-patterns.