Ingø Rübe

Co-Creator of the Hingepoint Framework, Co-Founder of Founder of @KILTProtocol, former CTO @burda_news and board member @drupal.

[Ingø Rübe @ingoruebe]

Good morning. I would like to give a quick updated after being mostly silent on social media for half a year:
In 2025 I supported galaniprojects GmbH with adapting their dev department for the AI-age. It took us 15 month in total to figure out the right balance between speed and accountability to deliver high quality software significantly faster and better than before. What was meant to be an internal project caught the attention of friends and business partners in early 2026. We were asked to share our experiences and make workshops for dev departments of companies of different sizes.
So we did the only thing reasonable: I teamed up with Anastasia Galani who has 25 years experience in Software Development Process Engineering and started to formalize our findings, put them into a framework, and make it available for anyone and any company open and free of charge.
We present the results at https://t.co/Y3gRgymveV and I would like you to check it out and potentially join our new community.

What we have in particular:

- The Tandem Manifesto (2 Pages, Creative Commons License) is a working agreement between Human Engineers and Coding Agents

- The Hingepoint Guide (37 pages, Creative Common License) explains how to operationalize Hingepoint in a corporate environment, including all processes, artifacts and responibilities.

- The LLM-agnostic SKILL and templates (Apache-2 open source license) instructs the coding agent to behave according to the working agreement and the processes humans learned from the guide.

I am pretty excited to share this.
Our success depends on people sharing it, improving it and using it. So please check it out, join the community, and share with all your dev friends, your CTO friends, and anyone else in software development.

[Ingø Rübe @ingoruebe]

I still have access to Mythos

[Ingø Rübe @ingoruebe]

@martinsparksdev Of course an entry in ERC-8004 does not make anything enforceable. What I would suggest is, using the Validation registry to store verifiable credentials, issued by third party entities, which make a claim against the Agent enforceable.

[Ingø Rübe @ingoruebe]

88% of organizations report AI agent security incidents, yet only 14% deploy with proper approvals. Meanwhile, ERC-8004 just went live on Ethereum mainnet. Are we building accountability infrastructure fast enough for the agentic wave already here?

[Ingø Rübe @ingoruebe]

@giottodf PoP ignores companies, DAOs, organisations, AI-Agents. That means it ignores reality. When 100 people make individual PoP and then form a DAO and act together then you have exactly what you had before :-)

[Ingø Rübe @ingoruebe]

Observation: every AI company talks about “alignment” for models.
But almost no one talks about alignment for agents even though agents actually take actions in the real world.
Model alignment is about outputs. Agent accountability is about consequences.
Two different problems but both critical.

[Ingø Rübe @ingoruebe]

the biggest bottleneck for AI agent adoption isn't capability. It's trust.
Agents are getting incredibly good at doing things. But we have no infrastructure for holding them accountable.
Capability without accountability is dangerous.

[Trustful Agents @TrustfulAgents]

AI agents will soon run critical workflows, sign contracts, and negotiate on your behalf. But today, there’s no reliable way to know they’ll do what they promise.

So why trust unaccountable AI agents? We need a real trust layer for the agentic web.
Stay tuned. We’re building it.

[Ingø Rübe @ingoruebe]

As AI agents start signing contracts and handling real economic activity, what should back their behavior? Reputation, insurance, collateral, something else? Who should be accountable? The user or the provider?

[Ingø Rübe @ingoruebe]

Adding Trust to Trustless Agents

[Trustful Agents @TrustfulAgents]

Hello 👋 We’re a new project focused on trust and accountability in the agentic web.

As #AI agents take on real work for real customers, and sometimes cause real damage, we’re building a collateral-based accountability model to for AI agents.

Stay tuned for more!

[Ingø Rübe @ingoruebe]

You wouldn’t hire an unlicensed contractor or trust an uninsured bank. So why are we handing critical tasks to AI agents with zero accountability?
Brand trust won’t scale as agents multiply. We need a real trust layer for AI agents.

[Ingø Rübe @ingoruebe]

AI agents will soon handle your critical business processes, sign contracts, and negotiate on your behalf.
But right now, there's zero way to know if an agent will actually do what it promises.
That's a problem we need to solve before it's too late.

[Ingø Rübe @ingoruebe]

@azaztrader01 We think of a number of neutral councils for different verticals (defi, medical, …) all with a number of reputable councillors. Councils vote after an evidence period, where claimers and providers discuss and hand in evidence and counter-evidence.

[Ingø Rübe @ingoruebe]

Great to hear that Filecoin is involved here. In February we will present a system for verifiable agent reputation through collateralized commitments and transparent dispute resolution. Of course based on ERC-8004. Would be nice to speak with Filecoin community on potential collaborations.

[Filecoin @Filecoin]

ERC-8004 trustless agents rely on verifiable identity.

The @ethereum agent builder community uses Filecoin Pin to store agent identity, reputation, and metadata onchain.

Agent0, an open source SDK by ERC-8004 author @marco_derossi, helps developers create these agents.

[Ingø Rübe @ingoruebe]

Happy to be part of it. Thanks for inviting me!

[Lucy @luce_lipped]

Tomorrow, Web3 takes the floor at the UK Parliament.🏛️

@AcademyPolkadot, @xcavateofficial, @gavofyork & others join a high-level debate on how education, innovation & opportunity can shape the UK’s role in the next era of the internet.

And I am hosting - what could go wrong? 😁

[Ingø Rübe @ingoruebe]

If you’re interested in AI agents or AI native companies, you might want to spend 20min listening to my talk at the web3summit im Berlin, which was recorded 2 weeks ago. Enjoy! Looking forward to comments, suggestions, ideas! https://t.co/nEqwWiwECM

[Ingø Rübe @ingoruebe]

@NattiCrypto @Polkadot @lolmcshizz @AcademyPolkadot @DaKampe @polkadotmkt @Web3summit Shoutout to the artist as well!

[Ingø Rübe @ingoruebe]

Jay is right ofc. But PoP really has surprisingly little use cases. General elections would be one. For blockchains airdrops come to mind. Anything else (like social media, voting in communities, making decisions) is rather a question of identification and accountability. Not of personhood. PoP does not only discriminate AI, it also discriminates companies, DAOs, NGOs and so on. It’s a backwards facing concept.

[Ingø Rübe @ingoruebe]

The Events Bounty is the lighthouse of Polkadot‘s OpenGov! Real professionals make funding decisions on behalf of the Polkadot community. Independent, ROI driven, result oriented and still friendly. It‘s one of the examples where decentralisation clearly works.

[Polkadot Events @PolkadotEB]

Welcome to the new official account of the @Polkadot Events Bounty 🎉🌍

After losing our old account (thanks, X 🫠), we’re back - new handle, same mission:

Funding and spotlighting Polkadot ecosystem events around the world.

Here’s everything you need to know 👇

[Ingø Rübe @ingoruebe]

Thrilled to speak at Web3 Summit in Berlin this year. If interested how we can bring a layer of accountability into AI agents, please join!

[Web3 Summit @Web3summit]

We’re pleased to announce our next speaker, @ingoruebe, founder of @Kiltprotocol

Grab your ticket here: https://t.co/cLKqb6TaLM