Re: [CG Report v1.6.2 draft] Vouch Protocol: Continuous State Verifiability for Autonomous AI Agents

Thanks Manu, the support means a lot.

Picking up the object-capabilities thread: you and Alan were both right
that the delegation layer is really an OCAP problem, not a VC one. I have
taken a first pass at the redesign off Alan's feedback. It drops the fixed
depth limit and moves to a proper attenuation rule, where each delegation
link must be a strict subset of its parent and can never broaden it, with
cost control handed to the verifier rather than baked into the protocol.
That puts the chain model on the OCAP and ZCAP footing rather than treating
delegation as another VC.

I have put it up as a draft for review here:
https://github.com/vouch-protocol/vouch/pull/42. Three edge cases are left
open as notes, where I would most value input.

Alan, if you have a moment, your read on those three would be really
helpful: whether a strict subset on any one dimension is tight enough,
whether the leaf and termination condition needs spelling out, and how
revocation should cascade down a chain. No rush, and thanks again for the
original catch.

Ram

On Mon, 8 Jun 2026 at 18:05, Manu Sporny <msporny@digitalbazaar.com> wrote:

> On Sun, May 31, 2026 at 5:41 PM Ramprasad G <ram@vouch-protocol.com>
> wrote:
> > I would like to draw the group's attention to a Community Group Report
> draft for Vouch Protocol, a continuous-state-verifiability layer for
> autonomous AI agents built entirely on existing W3C primitives.
>
> Just a quick note of support for the work that Ramprasad has done. I
> have had a chance to review it and the structure and architecture seem
> good, in that they re-use many of the primitives we've built in this
> group in ways that make sense.
>
> I'll also note that the Vouched.id folks might be interested in this
> work as well. I was speaking with Brent last week and he mentioned
> that there are something like 30+ separate specs proposing work on
> agentic identification and/or delegation at IETF. So, the area is hot
> right now.
>
> I agree with Alan's insights on delegation chains (as always), and I
> think there is room for the spec to align w/ many of the lessons
> learned around object capabilities over the past 25+ years. I've found
> Ramprasad smart, humble, and easy to work with... so I'm sure things
> will align as they mature.
>
> Just a general +1 to the work... it's headed in a good direction, IMHO.
>
> -- manu
>
> --
> Manu Sporny - https://www.linkedin.com/in/manusporny/
> Founder/CEO - Digital Bazaar, Inc.
> https://www.digitalbazaar.com/
>