Skip to content
View landjunge's full-sized avatar
🏠
Working from home
🏠
Working from home
  • 19:45 (UTC +02:00)

Block or report landjunge

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
landjunge/README.md

Daniel Filipek · landjunge

System Designer & Product Architect — building understandable, controllable systems with AI as a technical partner.

Public GitHub is only these six products plus this profile. Everything else is private.

Netzwerkpunkt

netzwerkpunkt.de is the public home for my projects. Each product has its own site and repository.

Project Product site Source Responsibility
4AllPass 4allpass.netzwerkpunkt.de GitHub Local-first password manager. Autofill, zero-knowledge, agents only after Allow/Deny.
Tollgate tollgate.netzwerkpunkt.de GitHub Independent safety, policy, budget and network-control layer for agents.
gnom-hub-v1 gnom-hub-v1.netzwerkpunkt.de GitHub Local multi-agent desk. Brainstorm freely; execute deliberately.
ThreadDesk threaddesk.netzwerkpunkt.de GitHub Persistent context, threads, files and decisions. Stores work; never executes it.
Versteh-Mir — GitHub Checks understanding (know / unsure / nothing) before anything is built.
Agent Authority Lab — GitHub Experimental composition monitor. Fake world only, no real secrets.

What I work on

I explore a practical question:

How can people use AI agents productively without losing context, control, credentials, or responsibility?

The answer is not one opaque super-system. I build separate tools with clear boundaries and clean interfaces. Each project must remain useful on its own.

These projects are related by principles, not fused into a monolith:

  • You decide — no silent autopilot.
  • Local first — as little cloud dependency as practical.
  • Clear responsibility — context, execution, credentials and action control stay separate.
  • Inspectable work — specifications, tests, reviews and attempts to break assumptions are welcome.

How I build

Product vision, priorities, system boundaries and architecture decisions come from me. AI is my technical partner for implementation, testing, research and documentation.

I am not presenting myself as a traditional software developer or security specialist. I focus on identifying problems, decomposing them into independent tools and defining how they should behave. The technical implementation must remain reviewable—especially when security is involved.

Honest project status

These are actively developed projects and experiments, not finished enterprise products.

4AllPass is security-critical and has not yet received an independent third-party audit. Tests, specifications and internal reviews are useful evidence, but they are not a substitute for an external audit. Do not store irreplaceable production secrets in it yet.

Deutsch

Ich entwickle Systeme und Produkte mit KI als technischem Partner. Meine Stärke liegt darin, Probleme zu erkennen, sie in eigenständige Werkzeuge zu zerlegen und klare Grenzen sowie Schnittstellen festzulegen. Vision, Prioritäten und Architekturentscheidungen kommen von mir; KI unterstützt bei Umsetzung, Tests und Dokumentation.

Alle Projekte sind über netzwerkpunkt.de und ihre eigenen Subdomains erreichbar. Die Werkzeuge bleiben eigenständig und können über saubere Schnittstellen zusammenspielen.

Ă–ffentlich auf GitHub sind nur die sechs Produkte in der Tabelle plus dieses Profil. Alles andere bleibt privat.

Ich bin kein klassischer Softwareentwickler und kein Security-Spezialist. Deshalb sollen Code, Annahmen und Sicherheitsgrenzen offen ĂĽberprĂĽfbar sein.

Feedback, Tests, Reviews und ehrliche Kritik sind willkommen.

Popular repositories Loading

  1. gnom-hub-v1 gnom-hub-v1 Public

    Local multi-agent control hub — brainstorm freely, execute only when you press Execute. Python · FastAPI · no Docker.

    Python 1

  2. tollgate tollgate Public

    Safety layer for AI agents — Protect · Route · Prove. Budgets, tool-loop hard stops, failover, chaos DR proof. OpenAI drop-in · MCP · n8n.

    Python

  3. landjunge landjunge Public

    Profile

  4. threaddesk threaddesk Public

    Local-first control layer: persistent threads for AI work. Stores context, never executes.

    Python

  5. 4AllPass 4AllPass Public

    Local-first password manager. Autofill. Zero-knowledge. Agents only after Allow/Deny. / Lokaler Passwort-Manager. Autofill. Zero-Knowledge. Agenten nur nach Allow/Deny.

    TypeScript

  6. agent-authority-lab agent-authority-lab Public

    Experimental security lab: deterministic composition monitor (Prototype v0.2). Fake world only, no real secrets.

    Python