[WIMSE] Re: IETF 127 Hackathon: WIMSE interop scenarios for agent call chains
Ben Le <ben.le@redthreadsec.com> Sun, 15 November 2026 00:17 UTC
Received: from mail-yw1-x112e.google.com (mail-yw1-x112e.google.com [IPv6:2607:f8b0:4864:20::112e]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange x25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by mx.ietf.org (Postfix) with ESMTPS id 2148A46 for <wimse@ietf.org>; Sun, 15 Nov 2026 00:17:18 +0000 (UTC)
Authentication-Results: mx.ietf.org; dkim=pass header.d=redthreadsec.com header.s=google header.b=iZepWYv3; spf=pass (mx.ietf.org: domain of ben.le@redthreadsec.com designates 2607:f8b0:4864:20::112e as permitted sender) smtp.mailfrom=ben.le@redthreadsec.com; dmarc=pass (policy=none) header.from=redthreadsec.com; arc=pass ("google.com:s=arc-20260327:i=1")
Received: by mail-yw1-x112e.google.com with SMTP id 00721157ae682-8b2f7573eafso1239157b3.2 for <wimse@ietf.org>; Sat, 14 Nov 2026 16:17:18 -0800 (PST)
ARC-Seal: i=1; a=rsa-sha256; t=1794701837; cv=none; d=google.com; s=arc-20260327; b=amVs9XXna9wjn3mhTSA+hHcd1uAnGAUHGA4AIcGkd+i0iwlgXI1HkQhBe9yA2Xyj5P c+MpiBjQeLay0DgZtxlvPwTvhLXu36AEWVpwHkVIn0YZOpqzOBW7MuaLklNT9ABXF7w/ wL7ZxvVD9w5xihuV7AWJHQxOjqtxk4K9nE3LWVNpCWBCfFPpqi9vbU1r1KSXT9kOaSi1 0jkFh2u5eQBPVWlSnaHH7blsu/gcUCoFZU7rWQroJf3VD2S9YWQCWNqJu/IXvVfPN85S 2tFson6dYsXrSjkIT+KrjozFEBF/YySFjw8Ub36FPcvD/pInxID6gpcnUca5rlM/msIh N4dg==
ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20260327; h=cc:to:subject:message-id:date:mime-version:references:in-reply-to :from:dkim-signature; bh=z5X9H+HeH171Wrec3Ezws2SU0OE2bc1GlMXOxzykhEg=; fh=4U5JOPmfsRSk7rHPZtzulgnFb42rTxleGvEAfALfPZU=; b=R2tGwx3XwFpbTIHrKtAhKkBAagbVe6AmnRHvNIIQGWcmGqi/ziHV4BxUTrFqJ7MD5C FRKOOJBwPBLY+2ThVG5BiQ2Rxnvf155F6mTPBI71aGY+Jri7KlPfJZgE0qrArFVmdwtm Yh0EZn/HLll0WVvuI0x9lWL1p2quwax+cObJS88ub+x1OIDC6d4e966Fr3Jxr5Jghe85 B8eCPoC+SRB+6wXQNGJ/LVnZugwi09qUOImWTilpOJfE7P1wJ0srufIIbYh4Az8LqA8i 3fIVyPtUs/BkrYAMn0GLhFo3o/hC+wQbwPuBNvLXVRaBcoPUW3SMjvYkP/vMadge1l7j Z55w==; darn=ietf.org
ARC-Authentication-Results: i=1; mx.google.com; arc=none
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redthreadsec.com; s=google; t=1794701837; x=1795306637; darn=ietf.org; h=content-type:cc:to:subject:message-id:date:mime-version:references :in-reply-to:from:from:to:cc:subject:date:message-id:reply-to :content-type; bh=z5X9H+HeH171Wrec3Ezws2SU0OE2bc1GlMXOxzykhEg=; b=iZepWYv3Xf0srBbKifddKE34tSdD9UrQ480taNNrL0Q4hSR0TondCF8MEcDxbtVVxJ ul5Op5qtZ2IEXPuGxxqzgEicrBhQXIy7RV4d4BVteNiZLE5VwaZXDDxddwhZ+2osfBed 7hYJJNSiXEsU2kaaXudGqIV7PLTb+pWt8k6LuoAstnA3UbLV8b1DMih6IcMmksszwCjO ArqQQshpwBeZuwrI4qwvSFIq7GGZFAVQSWd51n3P839la50d14rhMjscZ/cKY44m0YUa cknw2DJKn0KwsxNyqn/uUgtmCCYB4U2G4Scv6pcC3iZAQo+dcFm32NaLZnpcF3Sh4gbd FiVQ==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20260707; t=1794701837; x=1795306637; h=content-type:cc:to:subject:message-id:date:mime-version:references :in-reply-to:from:x-gm-gg:x-gm-message-state:from:to:cc:subject:date :message-id:reply-to:content-type; bh=z5X9H+HeH171Wrec3Ezws2SU0OE2bc1GlMXOxzykhEg=; b=0EmhPQT19t78J5M8C2u0VdW5C69LIv8CNvVk+xNc8ATa4Y/TNrYXH5vGULs5JES75T h0ha83WhMYFb578oQeFgi+s7j5u4PAsc1/wJyxLARxiXbCZSY2c0To0ILa1iRVBs11bd OGmZkK9+Z52ZG9gRY+pAygkPwv6AW6rdhuDo7dBofHRNUM/VR21mvLDRwoOQolsPNSYI Ya4WYvEl8RrExqQthA4ES2qlMzfw3jPPEmbSNtIxgplXI3a3mpwJNJSWeU/200H1HHye vH5gq08++iHlPFE4ntIG6uiTuOqv0utBNQ8vyjfsX+3jDH3QJ5M0ErVd/XhrtdqPB05/ X0xg==
X-Gm-Message-State: APuU4Z8llhGr8wwrmSqDGXfnD4dA1WFKEM76TsRtlSaLr+cgZzOAB7kU ZyjHcbZgzECtMKEk4s5esJltzi3MFOjU5RBiIl43Yhfe9QAhxCuA/Q3t2sw8e7hxmPom+wFFnJy h6GKkFpbNVnvlG2kqHvd99GfY/TkMh7EpABdicsQgoVQ=
X-Gm-Gg: AYBFou3tx6Fuvj5MAx4mYfm79gRk29NFv5FIRazewcGsPIoBp5PN36aI1DPb8Ky1PZK FhcBSR3OLedWQ+h2yOLtPoTLx9LS3lSVuDUX9Rqvp3+Lx127eMHbZc30gVLlWyd441wXzqFN8gH j6XKtSHWKHe3wl2Nx+mCj9cuJsj+wf9HhwWQ8Hvx21Q2q1bd+7XbGN52CMHN4vtJcjeS0l4qLhY 7DZCI1W3uDjttI962caJCfWKZAl53lwoApWtPGVKGwQt+9+C4qMOl20clz88aSYI+6q2IZXSS/S yMsh4MHVBWUHDbcGKf8EKb3gRRUfsjqy/aoSO8fRGvsLr6JTj9OqgCOp246nNogxR/2poYyxTby vW5jfpKs5SCmM/RI7Hs0EdTDyle05mM1xNShoQbJ/bse+Dfzp2ydq4HqyqG2fdzqZDSTjI1KBAU ry8VucJDRbOi5yR1Hxyhgak7jFRAc2heB+Y1Gaegi6ncb3Ob3Y5IK4IAS9X1uzfdT4hn0JLv3Qa iJ9lh5sjZ5RnuHxE3nz1Gc84h5u
X-Received: by 2002:a05:690c:e0c4:20b0:8ab:3f98:28d2 with SMTP id 00721157ae682-8b4a9f83f86mr49767147b3.0.1794701837413; Sat, 14 Nov 2026 16:17:17 -0800 (PST)
Received: from 101988054943 named unknown by gmailapi.google.com with HTTPREST; Sat, 14 Nov 2026 16:17:16 -0800
Received: from 101988054943 named unknown by gmailapi.google.com with HTTPREST; Sat, 14 Nov 2026 16:03:18 -0800
From: Ben Le <ben.le@redthreadsec.com>
In-Reply-To: <CAM7ZfzM+EqfDXq9EppDLFi4adQbrFq9u-QY3cYbS0OhBqBewkA@mail.gmail.com>
References: <CAM7ZfzM+EqfDXq9EppDLFi4adQbrFq9u-QY3cYbS0OhBqBewkA@mail.gmail.com>
MIME-Version: 1.0
Date: Sat, 14 Nov 2026 16:17:16 -0800
X-Gm-Features: AclHuK_SfmlQbPpRGctO6rlQN4bd5De7066kvfVDeNjaT1lWHUKQ9-iSJ74Hp5M
Message-ID: <CAO8m13dETof5ArvHGu+SXO8ab9U6yv_WOcggBOp1qrqud5pvxg@mail.gmail.com>
To: arun@advitlabs.com
Content-Type: multipart/related; boundary="0000000000007f88b7066045956c"
X-Spamd-Bar: -------------------
Message-ID-Hash: XVQET6XQTDK6EFJFDLBQE376GTUYQXYW
X-Message-ID-Hash: XVQET6XQTDK6EFJFDLBQE376GTUYQXYW
X-MailFrom: ben.le@redthreadsec.com
X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; loop; banned-address; emergency; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header
CC: wimse@ietf.org, team@emiliaprotocol.ai, marc@mintid.net, kieran@kierans.net
X-Mailman-Version: 3.3.10
Precedence: list
Subject: [WIMSE] Re: IETF 127 Hackathon: WIMSE interop scenarios for agent call chains
List-Id: WIMSE Workload Identity in Multi-Service Environment <wimse.ietf.org>
Archived-At: <https://mailarchive.ietf.org/arch/msg/wimse/nPDAthCFZZP1OsSsJPKmto0qNdY>
List-Archive: <https://mailarchive.ietf.org/arch/browse/wimse>
List-Help: <mailto:wimse-request@ietf.org?subject=help>
List-Owner: <mailto:wimse-owner@ietf.org>
List-Post: <mailto:wimse@ietf.org>
List-Subscribe: <mailto:wimse-join@ietf.org>
List-Unsubscribe: <mailto:wimse-leave@ietf.org>
Hi Arun, Thank you for the four details, and sorry for the wrong company name. It is OmniArx, and the P7 file now says so. I have added your inputs to the P7 case so any implementation can reproduce your result: - verify at time 1800000000, because the vectors use the fixed clock from OASNT Appendix A and a verifier that stops at the first failure never reaches the display check at today's clock; - the key from A.1 enrolled, the action from A.2 (V1), the request from A.5 (V4); - fresh replay state, because R2 has the same jti as V5 and is also refused as a replay after V5; - a verifier that holds the action. Without a request-to-action mapping the display digest is not recomputed (OASNT-ENFORCE Section 5.1), so accepting there is the expected result, not a failure. The file also records your result: V5 accepted, R2 refused with display-mismatch as the only reason. Our verifier does not hold an action or a request-to-action mapping, and our receipts do not carry what the person saw. So we will not run P7 ourselves. We will list it as an OmniArx result and not claim it for Redthread. P7 does show why the root receipt signed by the principal's authenticator, which Appendix B of the passport draft leaves open, matters: action binding alone would accept R2. I would like to talk through how an approval record could sit in the root receipt without moving the display check out of OASNT. Could we do that at the side meeting on Tuesday 17 November (18:45 to 19:45 PT, Plaza A)? I will send the updated P7 file with the next scenario update. Best regards, Ben Le Redthread Security On Sat, Nov 14, 2026 10:04 PM, Arun Thallapelly <arun@advitlabs.com> wrote: > Hi Ben, > > Thank you for adding P7. Four details decide whether two implementations > get the same result from it: > > - Verify at time 1800000000. The vectors use a fixed clock (OASNT Appendix > A). At today's clock the token is not yet valid, so a verifier that stops > at the first failure never reaches the display check. > - Inputs: the key in A.1 enrolled, the action in A.2 (V1), and the request > in A.5 (V4). > - Use fresh replay state. R2 has the same jti as V5, so after V5 it is > also refused as a replay. > - The case needs a verifier that holds the action. An enforcement point > without a request-to-action mapping does not recompute the display digest > (OASNT-ENFORCE Section 5.1), so accepting there is the expected result, not > a failure. > > With those inputs our verifier accepts V5 and refuses R2 with > display-mismatch as the only reason. > > One small fix: the company name is OmniArx. > > Best, > Arun > https://omniarx.ai > > On Sat, Nov 14, 2026 at 1:00 PM Ben Le <ben.le@redthreadsec.com> wrote: > >> Hi Alexander, >> >> As promised, the scenario files are attached: one JSON file per scenario, >> each with its setup, the expected decision, the reason to log, and whether >> any effect occurred. They cover the six I listed on 12 November, plus two >> from this thread: >> >> P1. Status list unreachable >> P2. Revocation after delegation, before effect >> P3. Delegation that widens >> P4. Cumulative limit across two services >> P5. Request rewritten in transit >> P6. Tool description or build changes between hops >> P7. Approval display differs from the executed action (Arun) >> P8. Final hop to a bearer-only third-party API (Kieran) >> >> Iman, both of your points are in. Confirmed revocation, an expired status >> cache and an unavailable status service have separate log reasons >> (status-revoked, status-stale, status-unavailable), even where the test >> policy denies all three. Every expectation is labelled with its basis: -00, >> -01 (planned, posted after IETF 127), a fixed test policy, or another >> draft. Where -00 left an unavailable status list to local policy, P1 states >> the fail-closed behaviour -01 adopts and labels it as -01. For the >> before-effect run, the expected result is yours: once the enforcing service >> sees the revocation, the protected operation produces no effect. >> >> Arun, P7 uses OASNT Appendix A.7.2 (R2) as the input, so any >> implementation can check its result. The passport draft lists a root >> receipt signed by the principal's authenticator as an open issue, and this >> result will inform it. >> >> Kieran, P8 has your three negative cases. I added one field to record: >> whether the bearer token had already been released when the revocation >> arrived. Once it has, the provider's token lifetime bounds the exposure, >> not the status list. >> >> Marc, your implementation fits P2, which records the time from revocation >> to the first refusal against the bound. On the network hours, Alexander's >> schedule decides. The hackathon starts at 09:30 Pacific on Saturday (17:30 >> UTC), so 17:00 UTC Saturday to 00:00 UTC Monday would cover both days. >> >> Results are very welcome at the side meeting on Tuesday 17 November >> (18:45 to 19:45 PT, Plaza A). >> >> Best regards, >> Ben Le >> Redthread Security >> >> On Sat, Nov 14, 2026 06:01 AM, Marc Miro i Rierola <marc@mintid.net> >> wrote: >> >>> Alexander, Iman, Ben, >>> >>> We can join remotely with a second implementation of the revoked-agent >>> case and Iman's before-effect result: the enforcing hop asks for a fresh >>> audience-bound proof, valid for 10 s, before the protected operation, which >>> takes effect only if that proof is accepted. It comes from an open-source >>> identity protocol whose status check is not a status list, which bears on >>> Iman's three reasons. >>> >>> The check is a zero-knowledge proof of non-revocation against a status >>> commitment that the issuer publishes every 30 s and that the verifier >>> accepts for at most 180 s on our test network. The verifier reads the >>> commitment; the issuer never sees the check. The three reasons come out >>> like this: >>> >>> - Confirmed revocation cannot be logged by the verifier, by design: it >>> never learns which credential it checks, so a revoked credential is refused >>> with the same code as a never-revoked one holding a witness for a >>> superseded commitment. The trace attributes the refusal with a control: the >>> issuer refuses the revoked credential's witness refresh, while a >>> never-revoked credential refreshes and is accepted again. Of the three, >>> this is the only one we can show on the test network. >>> - A commitment past its maximum age (the expired cache) is refused with >>> that same code, and the verifier's decision log names the commitment and >>> its age. A client cannot cause it on the test network: the issuer would >>> have to stop publishing. >>> - An unavailable status service is a refusal of its own (fail closed). >>> It can be shown only on a local stack; a record of one from 8 October 2026 >>> is in the source tree. >>> >>> A run on the test network on 13 November 2026 took 1 min 45 s, setup and >>> both paths included. Issuer revocation: last acceptance 19.7 s and first >>> refusal 25.2 s after the revocation, against a bound of 185 s. Holder >>> self-revocation: 14.6 s and 17.1 s after it was recorded, against a bound >>> of 245 s. That is one run; only the bounds are promised. >>> >>> We implement none of WIT, WPT or Txn-Tokens, so we would act at the hop >>> that takes effect: a sandbox issuer, revocation on both paths, and the >>> trace, which records every decision with its reason code and the time to >>> the first refusal (open source: https://gitlab.com/mintid/mintid, tag >>> public-v2026-11-13.1, test-harness/trace/revocation_trace.py). >>> Endpoints are given on request; the published packages carry the network >>> registry. Which UTC hours should the test network be online? >>> >>> Marc >>> MintID; speaking for myself. Offered as an IETF Contribution under BCP >>> 78 and BCP 79. >>> >>> >>> *Marc* >>> T.(+34)673482121 <+34%20673%2048%2021%2021> >>> mintid.net >>> >>>
- [WIMSE] IETF 127 Hackathon: WIMSE interop scenari… Alexander PELOV
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Iman Schrock
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Ben Le
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Iman Schrock
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Marc Miro i Rierola
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Ben Le
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Arun Thallapelly
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Ben Le
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Arun Thallapelly
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… Kieran Sweeney
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… ben.le
- [WIMSE] Re: IETF 127 Hackathon: WIMSE interop sce… ben.le