sandroandric

sandroandric / AgentHandover

Public

What if OpenClaw, Claude Code, Codex etc. knew how to do your work without you saying it? AgentHandover observes you, learns and teaches your agents with self-improving skills.

133
53
100% credibility
Found Mar 31, 2026 at 133 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

AgentHandover is a Mac menu bar app that observes your daily workflows to generate self-improving Skills executable by AI agents like Claude Code.

How It Works

1
📥 Download and install

Get the app from the website and double-click the installer to add it to your menu bar.

2
🔒 Quick setup

Grant screen access so it can safely watch your work, and download small AI helpers that run on your Mac.

3
Start teaching
⏺️
Record actively

Click record, do your task once, then stop to create a Skill right away.

👀
Observe passively

Keep working as usual while it spots repeating patterns over days.

4
✅ Review and approve

Check the menu bar for new Skills, tweak if needed, and approve them for your AI.

5
🔗 Connect your AI

One-click link to helpers like Claude so they can use your Skills.

🚀 Work smarter

Your AI now handles routine tasks exactly like you, getting better each time with feedback.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 133 to 133 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is AgentHandover?

AgentHandover is a Mac app (Rust daemon, SwiftUI menu bar, Python worker) that watches your screen, learns workflows through focus recordings or passive observation, and generates self-improving Skills—structured playbooks with steps, strategy, guardrails, and your voice—for AI agents like OpenClaw, Claude Code, or Codex. It solves the "tell agents everything" problem by turning your habits into executable handoffs via MCP server, REST API on localhost:9477, or CLI commands like `agenthandover connect openclaw`. Runs fully local with Ollama models like Qwen.

Why is it gaining traction?

Stands out from GitHub Copilot (what GitHub Copilot can do: code autocomplete, not desktop tasks) by capturing any workflow—like Reddit marketing or deployments—and feeding it to agents with feedback loops that refine Skills over executions. One-click agent pairing (OpenClaw what to do, what OpenClaw can do for me) and semantic search beat manual prompting; local privacy avoids Copilot IP concerns (what GitHub Copilot configuration needs for IP protection). Devs dig the zero-config handover to what OpenClaw AI is used for.

Who should use this?

Mac devs and ops folks automating repetitive desktop flows: marketers running Reddit community engagement (what is OpenClaw Reddit, what OpenClaw gateway), engineers handling staging deploys, finance teams filing expenses. Perfect for agenthandover users evaluating what GitHub pricing plan accesses company knowledge bases—this builds your own locally.

Verdict

Worth a spin for Mac agent enthusiasts—133 stars reflect early maturity, but solid docs, Homebrew install, and Apache 2.0 license (solid what GitHub license for personal projects) make it low-risk. 1.0% credibility score flags polish needed; productionize after more executions prove the loop.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.