Coding

Let your agent point at things on your screen with big arrows

  • Level Beginner
  • Setup~30 min (estimated)
  • Evidence Anecdotal
  • CostNot stated in source
  • No-codeNo
Claude CodeCodexSource-linkedBeginner-friendly

What they did

A macOS command-line tool plus a Claude Code/Codex skill lets the agent draw an arrow and a sign on top of any window to show you where to click. Clicks pass through and the arrow removes itself.

How it went

Open-source (MIT) with screenshots of real apps

Worth knowing

  • Open-source (MIT) with screenshots of real apps.

Prompt

Cleaned version rewritten from the source — not a direct quote.

Walk me through [setting up two-factor login in this app] step by step. At each step, draw a big arrow on my screen pointing at exactly what I should click, with a short label, and wait for me to say 'next'.

Recommended

Try it in Grok Bot

  1. Copy the prompt (button below).
  2. Open Grok Bot chat and paste it. Replace any [bracketed] placeholders.
  3. Connect the listed tools (as needed) as connectors when Grok Bot asks.
  4. If the task repeats, ask: "Make this a daily routine."

Alternatives: the same prompt works in Claude Code, Codex, Claude, ChatGPT Agent, OpenClaw.

Set it up

Stack: Claude Code, Codex

  1. Read the original source for full context.
  2. Connect the tools your agent needs.
  3. Paste the prompt and replace [bracketed] placeholders.
  4. Run once manually and check the output before scheduling it.

Generic suggested steps, not from the source.

Evidence

Open-source (MIT) with screenshots of real apps

Source: Web · Verified by us: not yet · Report broken or outdated

Related use cases

ContentX · 2026-01

Turn YouTube channels into a weekly EPUB magazine

Every week, pulls new videos from a list of YouTube channels, rewrites their transcripts as magazine articles, and emails them to you as a newsletter plus an EPUB file.

  • Level Intermediate
  • Setup1–3 hrs*
  • Evidence Strong evidence
Claude CodeSource-linked
CodingX · 2026-10

Hermes Agent Nearly Doubled Speech-to-Text Accuracy on Hard Words: 99 to 183 of 240

A user turned his 15K saved Wispr Flow dictations into a test set for the local Whisper model behind his Hermes Agent. The agent turned his Wispr dictionary into a hint list, diffed raw vs corrected text to find words Whisper keeps mishearing, made find-and-fix rules, and tested every rule against the full history. With no fine-tuning, dictionary words went from 99/240 to 183/240 on Wispr clips and 19/45 to 43/45 on sentences he read aloud. It now archives the data nightly and looks for new fixes weekly.

  • Level Advanced
  • Setup1–3 hrs*
  • Evidence Strong evidence
Hermes AgentSource-linked
ResearchX · 2026-03

Build a NotebookLM + Obsidian research vault

Claude Code sends a pile of videos, PDFs and articles through NotebookLM and turns the output into linked Obsidian notes covering scenarios, actors and probabilities.

  • Level Intermediate
  • Setup1–3 hrs*
  • Evidence Some evidence
Claude CodeSource-linked