15:I[981,["658","static/chunks/658-a698693f56d0116a.js","183","static/chunks/183-6beeeb801c274aaa.js","474","static/chunks/app/%5Btype%5D/layout-631c4730b83c8223.js"],"Posts"] 16:T4da,I built a GitHub Action that lets you deploy Docker stacks to a Portainer instance sitting behind a private network — without opening any ports to the internet. The action spins up an ephemeral Tailscale node during the CI run using OAuth (so it never needs a long-lived auth key), reaches your Portainer API over the tailnet, deploys or updates your stack, then immediately logs the node out on cleanup — even if the job fails. The problem I was trying to solve: I run Portainer on a home server and didn't want to expose port 9443 publicly or set up a reverse proxy just to enable CD from GitHub Actions. Tailscale already handled my VPN — I just needed the CI runner to join the tailnet temporarily. It supports stack create/update/delete, private registry auth (GHCR, Docker Hub, etc.), env var injection, MagicDNS hostnames, and auto-detects your Portainer endpoint if you only have one.

Marketplace: https://github.com/marketplace/actions/portainer-tailscale-d... Happy to answer questions — feedback welcome, especially if you use Portainer + Tailscale differently.17:T503,I've been experimenting with agent frameworks and noticed that many workflows reduce to a simple pattern:

- Send input - Wait for a pattern - Branch on the match

This is essentially the classic Unix expect model, but applied to LLM conversations.

So I built expectllm — a minimal pattern-matching conversation flow library (365 lines of code).

Example:

    from expectllm import Conversation

    c = Conversation()
    c.send("Review this code for security issues")
    c.expect(r"found (\d+) issues")

    if int(c.match.group(1)) > 0:
        c.send("Fix

On Luck, My Daughter

Actergy

I Made Claude and Codex Argue Until My Code Plan Was Good

Show HN: GitHub Action to deploy to Portainer over Tailscale (no open ports)

OpenClaw Made Easy and Secure

Procedural Tron

No Skill. No Taste

Backing up Kubernetes clusters with Plakar

Show HN: Generative Art Synthesizer

A.I. Did What Super-Specialist Doctors Could Not – Claude Opus 4.6

Dual Income, No Kids: What We Know About 'DINKs' in the U.S.

Show HN: Expectllm – "expect"-style pattern matching for LLM conversations

Why Boys Are Behind in Reading at Every Age

Aldus founder PageMaker pioneer Paul Brainerd, 1947-2026

China tests first megawatt-class flying wind turbine

ARC-AGI-3 Update

A category satisfying all but the smallness condition from Giraud's theorem

Cruise: West Is Best

"I Trust in TextEdit."

Breaking Mixed Boolean-Arithmetic Obfuscation in Real-World Applications

QuickShell: Sharing is caring about an RCE attack chain on Quick Share

Bikepacking 1,400KM Through the Arctic Night

Soft Contamination Means Benchmarks Test Shallow Generalization

A chatbot's worst enemy is page refresh

Knowledge Work After Superintelligent AI

HN: Marlvel.ai – Free AI pulse for fragmented mobile public data

Ruby Is the Best Language for Building AI Apps

Why Is the American Diet So Deadly?

Show HN: Paste any URL → get a product launch video

How to Create Jobs for the 1.2B New Workers

Show HN: I built a serverless API for automated website screenshots