Khaliq khaliqgant

Moved

The IAM roles and Pod Identity guide is now maintained at https://agentrelay.com/docs/self-hosting/iam-pod-identity.

This gist is retired and no longer updated. Use the docs: the scripts there are checked downloads (SHA256SUMS), and the steps include fixes made since this gist.

All self-hosting guides: https://agentrelay.com/docs/self-hosting

Moved

The EKS install guide (including bring-your-own AWS and Okta sign-in) is now maintained at https://agentrelay.com/docs/self-hosting/install-eks.

This gist is retired and no longer updated. Use the docs: the scripts there are checked downloads (SHA256SUMS), and the steps include fixes made since this gist.

All self-hosting guides: https://agentrelay.com/docs/self-hosting

Does this exist? (yes)

"My co-founder and I want, say, my Claude Code agent to communicate with his Codex agent so we can both have them run and prevent any breakages from entirely parallel work. I would also love to be able to 'clone' his Codex agent so I can finish off something he was working on. Does anything like this exist? How are teams parallelizing work using agents?"

We've built this as two separate tools:

A cloud agent reviews your uncommitted file—live

Your laptop publishes an unsafe refund.js. A reviewer in a fresh Daytona sandbox requests a fix. You change the same local file, and the reviewer approves it. Relayfile carries the file and review across machines; git never enters the loop.

laptop: refund.js ──> Relayfile ──> Daytona: reviewer
laptop: review.md &lt;── Relayfile &lt;── Daytona: reviewer

	/**
	* Two agents collaborate on a marketing plan for the Agent Relay SDK.
	*
	* Claude posts its pillars to #marketing, Codex receives them automatically
	* (the broker injects channel messages into subscribed agent PTYs) and
	* writes the final plan — no API key required.
	*
	* Run:
	* npm run build && npm run marketing-plan --workspace=packages/sdk
	*/

	git rev-list --objects --all \|
	git cat-file --batch-check='%(objecttype) %(objectname) %(objectsize) %(rest)' \|
	sed -n 's/^blob //p' \|
	sort --numeric-sort --key=2 \|
	cut -c 1-12,41- \|
	$(command -v gnumfmt \|\| echo numfmt) --field=2 --to=iec-i --suffix=B --padding=7 --round=nearest


	# source https://stackoverflow.com/questions/10622179/how-to-find-identify-large-commits-in-git-history

	# mount temp
	sudo mount -o size=10M,rw,nodev,nosuid -t tmpfs tmpfs /tmp

	# Resize root volume, find correct name by running lsblk
	sudo growpart /dev/nvme0n1 1
	sudo resize2fs /dev/root

	# Resize data volume
	sudo resize2fs /dev/nvme1n1

	npm install nango@npm:@nangohq/cli@0.0.1-2e8ad3b5da52183bc8d2644594adcc52e3d1ad83

	Installs the package nangohq/cli@0.0.1... under the alias nango

	"contains(github.event.head_commit.message, 'deploy!')"

	# https://stackoverflow.com/questions/63619329/github-action-get-commit-message