Plannotator

This gives you a visual UI to review AI agent plans before they run and annotate diffs after they execute. You get four annotation types (delete, insert, replace, comment) with line-level precision, plus image annotation for UI feedback. The real value is the feedback loop: mark up what needs to change, click send, and the agent gets structured input. It auto-saves approved plans to Obsidian or Bear with YAML frontmatter and tags, which is great if you're already living in those apps. Works over SSH and in devcontainers. The 88 GitHub stars and failed security audits from Gen Agent Trust Hub and Snyk suggest this is early and you should review the code before using it in production.

Install to Claude Code

npx -y skills add supercent-io/skills-template --skill plannotator --agent claude-code