5 GitHub Repos You Need

The five repos

The five from the video, with the exact names (a couple are easy to mishear), what each one does and the setup prompt to paste into Claude Code or Codex. A repo is just a project folder someone has shared on GitHub. Your AI agent can install it for you.

1. No AI Slop

Scans any piece of writing and strips out the patterns that make it sound AI-written (“It’s not X. It’s Y.”, “What nobody tells you is…”), while keeping your own voice. It lists what it changed.

Paste this into Claude Code or Codex:

Install the /no-ai-slop skill globally from https://github.com/petergyang/no-ai-slop

The catch: It’s a skill, so it lives inside your AI tool. Then type /no-ai-slop followed by your writing. It’s also available as a ChatGPT plugin.

github.com/petergyang/no-ai-slop →

2. Comp AI CRM

An open-source CRM built for AI agents to run. The agent reads your inbox, researches contacts, fills in company records and books its own follow-ups, so you’re not the one updating the pipeline.

This one is a big build, so start with a sanity check. Paste this into Claude Code:

Look at https://github.com/trycompai/crm and read the README. Before installing anything, tell me in plain English: 1. What I would need to run it myself (accounts, keys, software, hosting) 2. Roughly how long setup takes for someone who isn't a developer 3. What it would cost me each month 4. Whether it can work with the email and calendar I already use Don't install or sign up for anything yet. Wait for me to decide.

The catch: You host it yourself. The README asks for a database, Docker, a Google or Microsoft sign-in setup and an AI key. It’s a developer-level project, not a ten-minute install. The makers have a setup prompt on trycrm.ai once you’re ready.

github.com/trycompai/crm →

3. Video Use

Lets a coding agent edit your videos. Drop raw footage in a folder, tell Claude Code what you want and it cuts the ums and dead air, colour grades, adds subtitles and hands back a finished file. No timeline.

The makers’ own setup prompt. Paste it into Claude Code or Codex:

Set up https://github.com/browser-use/video-use for me. Read install.md first to install this repo, wire up ffmpeg, register the skill with whichever agent you're running under, and set up the ElevenLabs API key — ask me to paste it when you need it. Then read SKILL.md for daily usage, and always read helpers/ because that's where the editing scripts live. After install, don't transcribe anything on your own — just tell me it's ready and wait for me to drop footage into a folder.

The catch: It needs an ElevenLabs API key (for the transcription) and ffmpeg, which the agent installs. It shows you its plan and waits for your OK before editing.

github.com/browser-use/video-use →

4. SkillSpector (NVIDIA)

Checks a skill for malicious instructions, hidden scripts and data leaks before you install it, and gives it a 0 to 100 risk score.

Two terminal lines: install it once, then scan the skill folder you downloaded.

uv tool install git+https://github.com/NVIDIA/skillspector.git skillspector scan "/path/to/the-skill" --no-llm

Full walkthrough, including how to read the report: wrightmode.com/scanner.

The catch: A low score means nothing known was found, not that the skill is guaranteed safe. Run it before the skill goes into your skills folder.

github.com/NVIDIA/SkillSpector →

5. Phone Harness

Lets your agent control your actual phone: it sees the screen, taps, types and moves through apps for you. Nothing gets installed on the phone.

The makers’ own setup prompt. Paste it into Claude Code or Codex:

Set up phone-harness for me. Clone https://github.com/ShawnPana/phone-harness into ~/.phone-harness, read `install.md` first, install it so `phone-harness` is a command on my PATH, and register it as an agent skill named phone-harness using `phone-harness skill` as the body. Then read `onboarding.md` and walk me through it.

The catch: iPhone works through the Mac’s iPhone Mirroring window, so you need a Mac. Android works over a cable or Wi-Fi from Mac, Windows or Linux. It reads text on screen, not icons, and it can’t do Face ID. Try it on a spare phone or a throwaway app first, not your banking app.

github.com/ShawnPana/phone-harness →