Nur auf Englisch

Why wearable interfaces could be the next home for coding agents

Originally published on OpenAgentix News.

wearablesagentsopen-source
German only

Warum Wearables eine Zukunft für Coding-Agenten sein könnten

Ursprünglich erschienen bei OpenAgentix News.

wearablesagentsopen-sourcedeutsch
Nur auf Englisch

How I Optimise My Agents: Logs, Reports, Vetoes and Small Steps

Getting an agent to do something once is easy. Getting ten scheduled agents to keep doing useful things for months, without me babysitting them, is a differe...

agentsclaude-codeautomationobservability
German only

Wie ich meine Agenten optimiere: Logs, Berichte, Veto und kleine Schritte

Einen Agenten einmal etwas tun zu lassen, ist leicht. Zehn geplante Agenten über Monate hinweg nützliche Arbeit machen zu lassen, ohne dass ich sie ständig b...

agentsclaude-codeautomationobservabilitydeutsch
Nur auf Englisch

What's New in My Self-Built Dashboard: Bilingual UI, Safer Agents, Auto-Scans

A day after I built my own Portainer with agents, the dashboard has already changed a lot. Using a tool every day is the fastest way to find what is missing,...

claude-codeagentshomelabdockerdevopsdashboardsecurity
German only

Neu im selbstgebauten Dashboard: Zweisprachig, sicherere Agenten, Auto-Scans

Einen Tag nachdem ich mein eigenes Portainer samt Agenten gebaut habe, hat sich am Dashboard schon viel getan. Ein Werkzeug jeden Tag zu benutzen ist der sch...

claude-codeagentshomelabdockerdevopsdashboardsecuritydeutsch
Nur auf Englisch

Letting an AI Agent Run My Homelab — and Why That's Nothing Special

My homelab now has agents. One reads logs and explains what went wrong. One runs every 15 minutes and checks whether everything is healthy — and if not, trie...

claude-codeagentshomelabdevopssecurity
German only

Ein KI-Agent betreibt mein Homelab – und warum das nichts Besonderes ist

Mein Homelab hat jetzt Agenten. Einer liest Logs und erklärt, was schiefgelaufen ist. Einer läuft alle 15 Minuten, prüft, ob alles gesund ist – und versucht,...

claude-codeagentshomelabdevopssecuritydeutsch
Nur auf Englisch

I Built My Own Portainer — Plus Agents — in a Single Day With AI

My homelab is one Proxmox host, one Docker host and a growing pile of containers: a few services I build myself, a lot of public images, static sites, a data...

claude-codeagentshomelabdockerdevopsdashboard
German only

Mein eigenes Portainer – samt Agenten – an einem Tag mit KI gebaut

Mein Homelab besteht aus einem Proxmox-Host, einem Docker-Host und einem wachsenden Haufen Container: ein paar Dienste, die ich selbst baue, viele öffentlich...

claude-codeagentshomelabdockerdevopsdashboarddeutsch
Nur auf Englisch

Long-running AI agents: progress files, clean states, checkpoints

Long-running AI agents need memory that lives outside the model. When work spans many context windows, each new session starts blank, so the harness must han...

harnesspatternsexplainer
German only

Lang laufende KI-Agenten: Fortschrittsdateien, Zustände, Checkpoints

Lang laufende KI-Agenten brauchen ein Gedächtnis außerhalb des Modells. Erstreckt sich Arbeit über viele Kontextfenster, beginnt jede neue Sitzung leer. Der ...

harnesspatternsexplainerdeutsch
Nur auf Englisch

Why SemVer and Conventional Commits Matter Even More with AI-Assisted Coding

When agents write code, the number of commits goes up. Not by a little: an agent that works through a task in small steps can produce a dozen commits in the ...

semverconventional-commitsgitagentsengineeringreleases
German only

Warum SemVer und Conventional Commits beim AI-Assisted Coding noch wichtiger sind

Wenn Agenten Code schreiben, steigt die Zahl der Commits. Und zwar nicht ein bisschen: Ein Agent, der eine Aufgabe in kleinen Schritten abarbeitet, erzeugt e...

semverconventional-commitsgitagentsengineeringreleasesdeutsch
Nur auf Englisch

When are multi-agent systems worth their tokens?

Multi-agent systems are worth their tokens when the task is broad, parallelisable and valuable enough to pay for several agents’ worth of context, and when y...

multi-agentcostcomparison
German only

Wann sind Multi-Agenten-Systeme ihre Tokens wert?

Multi-Agenten-Systeme lohnen sich, wenn die Aufgabe breit und parallelisierbar ist, wertvoll genug, um den Kontext mehrerer Agenten zu bezahlen, und wenn sic...

multi-agentcostcomparisondeutsch
Nur auf Englisch

Human in the loop for AI agents that works: fewer approvals, better ones

Human in the loop for AI agents works when approvals are rare, specific and informed, and it fails when a person is asked to approve every tool call. The fir...

governancepolicyopinion
German only

Human in the Loop bei KI-Agenten, der funktioniert: weniger Freigaben, bessere

Human in the Loop bei KI-Agenten funktioniert, wenn Freigaben selten, konkret und informiert sind, und er scheitert, wenn eine Person jeden Werkzeugaufruf be...

governancepolicyopiniondeutsch
Nur auf Englisch

AI agent sandboxing: filesystem and network isolation together

AI agent sandboxing needs two boundaries, not one: a filesystem boundary that limits what the agent can read and write, and a network boundary that limits wh...

securitysandboxinghow-to
German only

AI Agent Sandboxing: Dateisystem- und Netzwerkisolation zusammen

AI Agent Sandboxing braucht zwei Grenzen statt einer: eine Dateisystemgrenze, die begrenzt, was der Agent lesen und schreiben darf, und eine Netzwerkgrenze, ...

securitysandboxinghow-todeutsch
Nur auf Englisch

Workflows or agents? Choosing how much autonomy a step needs

Workflows vs agents is not a choice you make once for a whole system. It is a decision you make for each step: how much freedom does this step need? The answ...

architecturecomparisonpatterns
German only

Workflows oder Agenten? Wie viel Autonomie ein Schritt braucht

Workflows oder Agenten ist keine Entscheidung, die man einmal für ein ganzes System trifft. Man trifft sie für jeden Schritt: Wie viel Freiheit braucht diese...

architecturecomparisonpatternsdeutsch
Nur auf Englisch

Agent evals 101: tasks, graders and transcripts

AI agent evaluation means running an agent on defined tasks, many times, and scoring what actually happened in the environment, not just what the agent said ...

evaluationqualityexplainer
German only

Agent-Evals 101: Aufgaben, Grader und Transkripte

Die Bewertung von KI-Agenten (AI agent evaluation) bedeutet, einen Agenten auf definierten Aufgaben mehrfach laufen zu lassen und zu bewerten, was in der Umg...

evaluationqualityexplainerdeutsch
Nur auf Englisch

Context engineering: the smallest set of tokens that does the job

Context engineering is the practice of deciding what goes into a model’s context window at each step of an agent run, and what stays out. A prompt is one inp...

contexttoolsexplainer
German only

Context Engineering: die kleinste Menge Tokens, die die Aufgabe erfüllt

Context Engineering heißt, in jedem Schritt eines Agentenlaufs bewusst zu entscheiden, was im Kontextfenster des Modells steht und was draußen bleibt. Ein Pr...

contexttoolsexplainerdeutsch
Nur auf Englisch

Secrets for agents never belong in the context window

Secrets management for AI agents comes down to one rule: a credential must never enter the model’s context window. Anything the model can read can end up in ...

operationssecretssecurity
German only

Geheimnisse gehören nie in das Kontextfenster eines Agenten

Secrets Management für KI-Agenten lässt sich auf eine Regel verkürzen: Ein Zugangsdatum darf nie in das Kontextfenster des Modells gelangen. Alles, was das M...

operationssecretssecuritydeutsch
Nur auf Englisch

awesome-opencode: 108 Agents, 15 Skills, and a 2-Minute Setup for OpenCode

Setting up OpenCode properly – agents, skills, model configuration, MCP servers, project rules – takes about 30 minutes if you read the docs and do it by han...

weisser-dev/awesome-opencode

opencodeai-agentai-agentscliopen-sourceautomationmcp
Nur auf Englisch

opencode-remote-telegram: Control OpenCode from Your Phone via Telegram

opencode-remote-telegram is a Telegram bot that lets you control OpenCode from your phone. Pick a project, pick a model, type what you want — the agent works...

weisser-dev/opencode-remote-telegram

opencodetelegrambotaivibe-codingtypescript
Nur auf Englisch

Accidentally Committed Credentials to a Public Repo?

Step-by-step remediation for leaked API keys, tokens, and passwords in public Git repositories. What to do in the first 60 seconds, how to clean the history,...

gitsecuritycredentialssecretsincident-response
Nur auf Englisch

My Typical Frontend Hosting Setup: Strato + Cloudflare + Pages

Every frontend project I ship — agentic-ai.weisser.dev, scrumbuddy.org, cv.weisser.dev and others — follows the same hosting pattern: domain registered at St...

cloudflarehostingfrontendstratovitestatic-site
Nur auf Englisch

Building an Interactive AI Agent Workshop with Vite and Vanilla JS

A few weeks ago I set out to build a complete, interactive workshop about AI Agents, LLMs, and AI-assisted coding — something that could be used both as a li...

weisser-dev/agentic-ai-workshop

aiagentsopencodeworkshopjavascriptvite
Nur auf Englisch

OpenPastebin: A Self-Hosted Pastebin with Automatic Expiration

Every developer has reached for a pastebin at some point — to share a log snippet, pass a config file to a colleague, or dump some output during a debugging ...

weisser-dev/openpastebin

open-sourcepastebinjavascriptself-hosted
Nur auf Englisch

Writing tools agents can actually use: names, schemas and errors

Tool design for AI agents is interface design for an unusual reader: a language model that sees only the tool’s name, its description, its input schema and w...

toolsmcphow-to
German only

Tools schreiben, die Agenten wirklich nutzen können: Namen, Schemas und Fehler

Tool-Design für KI-Agenten ist Schnittstellen-Design für einen ungewöhnlichen Leser: ein Sprachmodell, das nur den Namen des Tools, seine Beschreibung, sein ...

toolsmcphow-todeutsch
Nur auf Englisch

The lethal trifecta: private data, untrusted content and a way out

The lethal trifecta for AI agents is a combination of three capabilities in one agent: access to private data, exposure to untrusted content, and a way to co...

securityprompt-injectionleast-privilege
German only

Die tödliche Dreifaltigkeit: private Daten, nicht vertrauenswürdige Inhalte und ein Weg nach draußen

Die „Lethal Trifecta“ (tödliche Dreifaltigkeit) bei KI-Agenten ist die Kombination dreier Fähigkeiten in einem Agenten: Zugriff auf private Daten, Kontakt mi...

securityprompt-injectionleast-privilegedeutsch
Nur auf Englisch

Building Custom Sub-Agents for AI-Assisted Workflows

As AI coding agents have become part of my daily workflow, I’ve started building specialized sub-agents — focused, single-purpose agents that can be invoked ...

shellai-agentsautomationopencode
Nur auf Englisch

OpenClaw Setup on VPS

Complete guide to installing and configuring OpenClaw on a VPS – Node.js 22, daemon setup, Nginx reverse proxy and firewall configuration.

openclawvpssetupdeploymentnginxlinux
Nur auf Englisch

OpenClaw Subagents Guide

Complete guide to using OpenClaw subagents for specialized tasks and multi-agent orchestration – architecture, routing, lifecycle and best practices.

openclawsubagentsmulti-agentorchestrationai-agents
Nur auf Englisch

OpenClaw with Qwen (Free Models)

How to use OpenClaw with Qwen Portal's free AI models using the qwen-portal-auth skill – 128k context, vision support, zero cost.

openclawqwenfree-modelssetuprag
Nur auf Englisch

OpenClaw Model Fallback Configuration

How to configure automatic model fallback in OpenClaw – auth profile rotation, cooldown progression, multi-provider chains and cost optimization strategies.

openclawmodelsfallbackconfigurationcost-optimization
Nur auf Englisch

Extending the Health Tracker: Data Visualization and Trends

Once you have a health data aggregation layer working, the next question is how to display it usefully. Raw numbers in a table answer “what happened” but don...

javascripthealthdata-visualizationpersonal-project
Nur auf Englisch

WW Food Tracker: Multimodal AI Agents for Nutrition Logging

Manual food tracking is tedious. You have to look up every ingredient, estimate portion sizes, and enter everything by hand — which is exactly why most peopl...

weisser-dev/ww-food-tracker

pythonai-agentcomputer-visionautomation
Nur auf Englisch

Building an AI-Powered Health Tracker

Health data is everywhere and nowhere at the same time. Fitness apps track your workouts, step counters track your movement, nutrition apps track your food, ...

javascriptaihealth-trackingpersonal-project
Nur auf Englisch

Agent Lyfta: An AI Personal Trainer Powered by Your Workout Data

Lyfta is a solid workout tracking app, but like most fitness apps, it keeps your data locked behind a mobile UI. The moment I found out there was an API, I w...

weisser-dev/agent-lyfta-api

pythonai-agentfitnessautomation
Nur auf Englisch

Forking OpenCode: How Terminal-Based AI Coding Agents Work

OpenCode is an open-source AI coding agent that runs in your terminal. Unlike IDE plugins or browser-based tools, it operates directly on your filesystem, re...

weisser-dev/opencode

opencodeai-agenttypescriptopen-source
Nur auf Englisch

Rebuilding My Personal Site as a Configurable Digital Card

Personal websites have a tendency to get stale. You build something, it looks good for a while, and then a year later the tech stack feels dated, the content...

weisser-dev/weisser.dev

typescriptreactpersonal-siteportfolio
Nur auf Englisch

A Multilingual, Configurable Resume App in TypeScript

There’s a category of developer who has built their own resume app, and I’ve become one of them. The practical motivation was concrete: I work in a bilingual...

typescriptreactresumemultilingual
Nur auf Englisch

My Server Essentials Setup: Automating a Fresh VPS

Every time I provision a new VPS, there’s a predictable set of tasks: update the package list, install essential tools, harden SSH, set up a firewall, config...

linuxserverautomationinfrastructure
Nur auf Englisch

Connecting Open WebUI to AWS Bedrock Knowledge Base for RAG

AWS Bedrock Knowledge Base is Amazon’s managed RAG (Retrieval Augmented Generation) service. You point it at an S3 bucket full of documents, it handles chunk...

weisser-dev/AWS-Bedrock-Knowledge-Base-Function

awsbedrockpythonopen-webuirag
Nur auf Englisch

Codex Dart: Building a Game 100% with OpenAI Codex

Back in mid-2025, before AI coding agents became a routine part of my workflow, I ran an experiment: build a complete dart game from scratch using nothing bu...

weisser-dev/codex-dart

aicodexjavascriptexperiment
Nur auf Englisch

Codex Examples: Experiments and Learnings with OpenAI Codex

Before building full projects with AI assistance, I spent time systematically testing what OpenAI Codex could and couldn’t do. The codex-examples repository ...

weisser-dev/codex-examples

aicodexjavascriptexperiment
Nur auf Englisch

Bedrock Access Gateway: OpenAI-Compatible APIs for Amazon Bedrock

One of the most practical problems when adopting Amazon Bedrock is that the ecosystem around it assumes you’re using the AWS SDK directly. Most open-source t...

weisser-dev/bedrock-access-gateway

awsbedrockpythonopenaiapi-gateway
Nur auf Englisch

Contributing to Open WebUI Documentation

Open WebUI’s documentation had some gaps, particularly around the Pipelines framework and the process of building custom integrations. Since I’d been working...

weisser-dev/open-webui-docs

open-webuidocumentation
Nur auf Englisch

Self-Hosting Open WebUI: Your Own AI Workspace

Open WebUI is a self-hosted web interface for interacting with language models — think ChatGPT, but running on your own infrastructure, against models you co...

weisser-dev/open-webui

open-webuiself-hostedaijavascript
Nur auf Englisch

qrcode-gen: The Value of Building Your Own Simple Tools

QR code generators exist by the dozens. Search engines return pages of them. So why build another one? The honest answer is that I wanted one I trusted — no ...

weisser-dev/qrcode-gen

htmlqrcodetool
Nur auf Englisch

iOS QuickDraw: Machine Learning Doodle Recognition in React

Google’s Quick Draw is one of the most publicly accessible machine learning demos ever made — you doodle something, a neural network guesses what it is in re...

weisser-dev/iOS-quickdraw-react

typescriptreactiosgameml
Nur auf Englisch

My Digital Card: A Configurable React Business Card Component

Physical business cards are on their way out. Digital ones are still catching up, but most solutions are either SaaS products with subscription fees and vend...

weisser-dev/my-digital-card

typescriptreactportfoliocard
Nur auf Englisch

A Fully Responsive Multilingual Resume/CV as a React App

Maintaining a resume is one of those tasks that’s easy to put off because updating a Word document or a PDF template is tedious and the result is never quite...

weisser-dev/responsive-resume-cv-react

typescriptreactresumemultilingual
Nur auf Englisch

Building a Flexible Prioritization Table Because I Couldn't Find One I Liked

Prioritization frameworks are everywhere — MoSCoW, RICE, weighted scoring, Kano model. What’s less common is a tool that lets you define your own criteria an...

weisser-dev/flexible-priorization-table

javascripttoolproductivity
Nur auf Englisch

endless.beer: The Most Important Project I Ever Built

In 2020, I built endless.beer — a webpage that shows a beer being poured in ASCII art, endlessly, forever. There is no other functionality. It does not stop....

weisser-dev/endless.beer

javascriptfunasciihumor
Nur auf Englisch

Wohnflächenberechnung: Automating German Living Area Calculations

In Germany, calculating the legally recognized living area (Wohnfläche) of a property is not as simple as measuring rooms and adding up the square meters. Th...

weisser-dev/wohnflaechenberechnung

javascripttoolgermanyreal-estate
Nur auf Englisch

Building a Browser Typing Game Over a Weekend

Typing speed matters more for developers than most people admit. The bottleneck in programming isn’t thinking time — it’s rarely typing speed either — but th...

weisser-dev/typinggame

javascriptgametyping
Nur auf Englisch

Angular Docker: Filling the Gap with Automated Image Builds

For a long time, there was no official Docker image for Angular development. If you wanted a consistent Angular build environment in CI or needed to containe...

weisser-dev/angular-docker

dockerangularciautomation
Nur auf Englisch

DevTools Coverage Extractor: Finding Unused JavaScript in Production

Chrome DevTools has a Coverage panel that tells you which lines of CSS and JavaScript are actually executed when a page loads. It’s a useful diagnostic tool ...

weisser-dev/devtools-coverage-extractor

javascriptdevtoolscoveragetooling
Nur auf Englisch

Running GitHub Actions on Your Own Infrastructure with Docker

GitHub-hosted Actions runners are convenient but come with limitations: you get a fresh machine each time but no persistent state, limited compute for intens...

weisser-dev/github-actions-runner-docker

dockergithub-actionsciself-hosted
Nur auf Englisch

ELK Stack with Docker Compose: Practical Logging and Observability

The ELK stack — Elasticsearch, Logstash, and Kibana — is a standard combination for centralized logging and observability. Adding Beats (Filebeat, Metricbeat...

weisser-dev/elastic-docker

dockerelasticsearchkibanaelkmonitoring
Nur auf Englisch

TypeScript Blog: An Early Experiment Before Jekyll

Before settling on a Jekyll-based approach for this blog, I experimented with building a TypeScript-powered blog from scratch. The experiment lived in the ty...

weisser-dev/typescript-blog

typescriptblogcss
Nur auf Englisch

Node.js Microservice Blueprint with Express and MySQL

This is the Node.js counterpart to the Java microservice blueprint — the same REST API specification implemented with Express and a MySQL backend, designed t...

weisser-dev/microservice_blueprint_nodejs

node.jsmicroservicesexpressmysqlrest-api
Nur auf Englisch

Lightweight Java Microservices Without Spring: Embedded Jetty Blueprint

Spring Boot is the default choice for Java REST APIs, and for many projects it’s the right choice. But Spring brings significant startup time, memory overhea...

weisser-dev/microservice_blueprint_java

javamicroservicesjettygradlerest-api
Nur auf Englisch

Terraform on GCP: Spinning Up a Kubernetes Cluster as Code

The terraform-cloud-application repository contains the infrastructure code for the cloud application tutorial series — Terraform HCL files that provision a ...

weisser-dev/terraform-cloud-application

terraformgcpkubernetesinfrastructure-as-code
Nur auf Englisch

Cloud Application #3 Infrastructre as Code - Create K8 Cluster on GCloud with Terraform & Deploy Container

You’re thinking about your last game of Buzzword Bingo when you read the headline? I trust you immediately! But here I explain to you - what exactly is “Infr...

operationsmicroservicescloudterraformdockerkubernetes
Nur auf Englisch

Cloud Application #2 Dockerize our Services

Prerequisites

developmentdockercontainercloudmicroservices
Nur auf Englisch

Very simple & fast Node.js Deployment

At the moment I’m in the process of rebuilding everything here. My private laptop, the blog etc..

operationsnode.jspm2
Nur auf Englisch

HOW TO: Fedora looks & feels like Mac OS X Mojave

Introduction In this tutorial I will show you how to get some features and the layout of OS X on your Fedora.

linuxfedora
Nur auf Englisch

Cloud Application #1 B) Build your own Java Microservice

Prerequisites installed java 8 installed gradle basic knowledge of java and gradle ide with gradle support (i use eclipse) enjoy programming

devjavacloud applicationcloud
Nur auf Englisch

TradingView Pine Script: Forecasting Bitcoin with Custom Indicators

Pine Script is TradingView’s proprietary scripting language for writing custom technical indicators and strategies. It’s a domain-specific language that runs...

weisser-dev/tradingview_scripts

tradingbitcoinpinescripttradingview
Nur auf Englisch

Cloud Application #1 A) Build your own Node.js Microservice

Prerequisites installed Node.js and Npm basic knowledge of JavaScript and Node.js enjoy programming

devnode.jscloud applicationcloud
Nur auf Englisch

Build a Facebook Messenger Bot (Node.js)

🤖 Build your own Messenger-Bot

developmentnode.js
Nur auf Englisch

Build a simple WebCrawler (Python)

Prerequisites To complete this tutorial, you’ll need a local development environment for Python 3. You can follow this Tutorial: “How To Install and Set Up a...

developmentpython
Nur auf Englisch

PM2 - Advanced Node.js process manager

Anyone who works with node. js knows how simple it is to build small web applications. But the fact that even complex pages are possible shows us e. g. Airbn...

operationsnode.js
Nur auf Englisch

Wiki Crawler: Indexing Wikipedia into Elasticsearch with Scrapy

The wiki-crawler project combines two complementary technologies: Scrapy for web crawling and data extraction, and Elasticsearch for indexing and search. Sta...

weisser-dev/wiki-crawler

pythonscrapyelasticsearchweb-crawler
Nur auf Englisch

Searchify: A Facebook Messenger Bot That Proxies Google Search

In 2018, chatbots were still a genuinely new interface paradigm for many users. Facebook Messenger’s bot platform had just opened up, and there was a real qu...

weisser-dev/searchify

javascriptnode.jsmessenger-botchatbot
Nur auf Englisch

Generating Custom .onion Addresses with Shallot

Tor hidden service addresses are generated from RSA key pairs: the address is derived from the public key, which means you can’t choose an arbitrary address ...

weisser-dev/OnionAddressGenerator

shelltoronionprivacy
Nur auf Englisch

Host your own VCS Server

Why should I host my own Version Control System?

operationsvcs
Nur auf Englisch

Create a custom *.onion URL with Shallot

It is very simple to create your own domain for the Tor network. -> If you don’t know how to install the Tor service and host a site there, read the follo...

operationsshell
Nur auf Englisch

Host a Tor Hidden Service (Nginx)

Create and Publish a new Website on Tor

operationsnginx
Nur auf Englisch

Build a Game - 2048 (Java)

Currently only the game sources are available. I developed it in 2014. If I have enough time, I will refactor it and write a HOW TO for it. Until then you ca...

developmentjava