agentation

Originally frombenjitaylor/agentation
Installation
Summary

Human-in-the-loop UI feedback bridge — humans click elements, agents receive CSS selectors and component trees to fix code.

  • Drop a React toolbar into any app; humans annotate UI elements with feedback while agents receive structured CSS selectors, bounding boxes, and React component paths for precise code targeting
  • Supports two modes: copy-paste markdown for quick feedback, or MCP watch-loop for iterative agent-driven fixes with acknowledgment and resolution tracking
  • Integrates with Claude Code, Codex, Gemini CLI, OpenCode, Cursor, and Windsurf via universal MCP server setup or platform-specific config files
  • Includes HTTP REST API (port 4747), SSE event streaming, and webhook delivery for custom integrations; works offline with local session storage and auto-syncs when connected to server
  • Pairs with agent-browser for autonomous self-driving UI critique cycles and integrates into jeo skill's VERIFY_UI phase for structured evaluation workflows
SKILL.md

agentation — Visual UI Feedback Bridge for AI Agents

The missing link between human eyes and agent code.

Instead of describing "the blue button in the sidebar," you hand the agent .sidebar > button.primary. It can grep for that directly.


When to use this skill

  • Human needs to point at a UI element and give feedback — without writing selectors
  • Running iterative UI/UX review cycles between human and coding agent
  • Building a watch-loop where agent auto-fixes every annotation a human leaves
  • Capturing CSS selectors, bounding boxes, and React component trees for precise code targeting
  • Autonomous design critique via agent-browser + self-driving pattern
  • Integrating visual feedback into agent hooks so annotations auto-appear in agent context

Related skills

More from supercent-io/skills-template

Installs
5.4K
GitHub Stars
88
First Seen
Mar 4, 2026