Docs — Antelog
Set up and use Antelog
Antelog checks every action your AI coding tool wants to take, before it runs. It blocks six groups of disasters: leaking your data and destroying it. Everything else runs, and all of it is written to a log on your computer.
Get started
Install
Paste one command into a terminal on the computer you want to protect.
macOS and Linux
$ curl -fsSL https://github.com/YashSarwaiya/antelog-releases/releases/latest/download/install.sh | sh
Windows (PowerShell)
$ irm https://github.com/YashSarwaiya/antelog-releases/releases/latest/download/install.ps1 | iex
It needs no admin rights and starts Antelog in the background. Every download is checked against its published checksum. All versions are on the releases page.
Then restart your open agent sessions. Sessions that were already running aren’t protected until you do.
Check it works:
$ antelog doctor
Connect your tools
The installer connects every AI coding tool it finds: Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Qwen Code, Cline, OpenCode and Kilo Code. Antelog guards the tool, not the model, so any model you use inside them is covered.
Codex: type /hooks and approve Antelog’s hooks. Do it again after each Antelog update.
Installed a tool after Antelog?
Claude Code or Codex: run the install command again. Others: run the command below, then restart the tool. To take Antelog out of a tool, add --remove.
Cursor
antelog setup cursor
GitHub Copilot
antelog setup copilot
Gemini CLI
antelog setup gemini
Qwen Code
antelog setup qwen
Cline
antelog setup cline
OpenCode
antelog setup opencode
Kilo Code
antelog setup kilo
What it stops
Six groups of disasters
Antelog blocks these before they run. Leaks first, then destruction:
Leak
- Prompt injection: A web page, email or chat tells the AI to send data out. Antelog blocks it. Once the AI has read outside content (a web page, search results, a mail, chat or GitHub tool, or code cloned from a stranger), sending files, data, messages or code to a place that isn’t yours is blocked. Your company places are fine, and so is a place you named in your own request, unless the AI read from that same site in this session.
- Secrets: AI tries to show or send your keys. Blocked on the spot. Printing .env files, private keys, or the credential files cloud tools save after login. Printing a whole environment, a secrets manager, or a platform’s variables. Sending a password or key to a server Antelog doesn’t know; well-known services are always fine.
- Personal info: AI tries to send customer data outside. Blocked before it leaves. After the AI has read outside content: customer emails or copied records.
Destroy
- Data and backups: AI tries to wipe live data or its backups. Antelog stops it first. Commands and actions that aim to delete data are blocked.
- Cloud: AI tries to tear down live servers. Blocked before anything breaks.
- Code and files: AI tries to erase your work. It never goes through.
Blocked too
- Turn off the guard: Changing Antelog’s settings or log, stopping it, uninstalling it, or starting another agent with its checks off.
- Download and run: Downloading a script from the internet and running it immediately.
- Unreadable scripts: Running a script that is too large to check safely.
Antelog also reads the scripts a command runs and stops them if they do any of the above.
Never stopped: reading, building, testing and installing; writing and editing code; local and test databases; deploys.
Prompt injection
Hidden text in a web page, email or chat can tell an AI to send your data away. After the AI has read outside content, sending files, data, messages or code to a place that isn’t yours is blocked.
When something is stopped
- In every tool: the action is blocked and the agent is told why in plain words.
- If Antelog fails: in some cases, the action is blocked, not let through.
The AI check
The AI never stops anything. When an action looks risky, it writes a note in your log and the action runs. It can undo one kind of false alarm.
Use it
Commands
antelog doctor
antelog log
antelog blocked
antelog sent
antelog privacy
antelog rules
antelog verify
antelog dashboard
antelog report
antelog check
Your own rules
Put one rule per line in ~/.antelog/rules.txt. Write the command in plain words. Use * for anything in the same command.
# my rules
helm uninstall
git push * main
regex: drop\s+user
Your rules can only add blocks; they never turn off the built-in ones.
Settings
All in ~/.antelog/. Your agent can’t change them. Restart Antelog after a change; antelog doctor shows how.
rules.txt
allowed-servers.txt
hosted-ai-check
anthropic-api-key
What left your computer
antelog sent lists the files, data, code and messages your agent sent out, and where they went. Then it counts searches and web pages by site.
Tamper-evident log
Antelog writes every action to its log before it runs. Each entry stores the fingerprint of the one before it. If an entry is changed or deleted, antelog verify shows where.
Silent backup
Before an agent throws away uncommitted work, Antelog saves it first.
List backups:
$ git for-each-ref refs/antelog/backups
Bring one back:
$ git restore --source=refs/antelog/backups/<time> --worktree -- .
Team
Free includes an online dashboard for one person, with 14 days of history. Team is for companies: admins invite developers and set company rules and company places for every computer.
Limits
- Commands you run yourself, and agents Antelog isn’t connected to.
- Most deletes written inside code.
- A force-push that doesn’t name its branch.
Uninstall
Run it yourself in a terminal; your agent can’t uninstall Antelog. Your history is kept unless you add --purge.