Hi all,

Cross-posting to wimse@ and spice@: Actor Chain uses workload/actor
identity inputs and includes subset-disclosure profiles relevant to
selective disclosure.

"Actor Chain" draft version 01 available at: draft-mw-oauth-actor-chain
<https://datatracker.ietf.org/doc/draft-mw-oauth-actor-chain/>. It
continues the earlier draft-mw-spice-actor-chain
<https://datatracker.ietf.org/doc/draft-mw-spice-actor-chain/> work, now
scoped as OAuth because it defines RFC 8693 Token Exchange profiles for
actor-path continuity in multi-hop service, workload, and agentic workflows.

Feedback welcome, especially on the core substrate, disclosure profiles,
and what belongs in the core spec versus companion work. Additional
technical notes are included below as an aid to reviewers.

Thanks,
Prasad
on behalf of the co-authors
------------------------------

*Additional notes on Actor Chain (draft-mw-oauth-actor-chain
<https://datatracker.ietf.org/doc/draft-mw-oauth-actor-chain/>) for
reviewers:*

*What it is:* RFC 8693 defines Token Exchange and the "act" claim, but
leaves append-only actor-path progression, workflow correlation,
profile-controlled disclosure, and proof-bound hop continuity unspecified.
Actor Chain preserves the RFC 8693 meanings of "sub", "act", and "may_act";
adds profile semantics only when "actp" is present; and keeps ordinary
tokens compact. The core claims are "actp" (profile selector), "acti"
(workflow instance identifier), RFC 8693 "act" (the disclosed actor path),
and "actc" (compact cumulative commitment state in verified profiles). The
profiles avoid one-size-fits-all behavior: "declared" means AS-asserted
continuity, while "verified" adds actor-signed step proofs and cumulative
commitment state, and the disclosure modes are "full" for the full visible
actor path, "subset" for an ordered subset selected by the AS for that
recipient, and "actor-only" for only the current actor.

*What it is not:* Actor Chain keeps the OAuth token deliberately slim and
limited to actor-path state, continuity evidence, and receiver
acknowledgment. Other data-plane, actor-domain, or deployment-specific
information can layer on top or be evaluated by the AS, RS, gateway, or
resource-side policy engine. These include authorization policy and
decisions, application result provenance, terminal-result evidence, and
audit or compliance evidence.

*Related work:* draft-ietf-oauth-identity-chaining
<https://datatracker.ietf.org/doc/draft-ietf-oauth-identity-chaining/> and
draft-ietf-oauth-transaction-tokens
<https://datatracker.ietf.org/doc/draft-ietf-oauth-transaction-tokens/> are
complementary, providing cross-AS identity and authorization continuity,
and transaction/request context, respectively. Our companion SPICE drafts,
draft-mw-spice-intent-chain
<https://datatracker.ietf.org/doc/draft-mw-spice-intent-chain/> and
draft-mw-spice-inference-chain
<https://datatracker.ietf.org/doc/draft-mw-spice-inference-chain/>, extend
this separation: Intent Chain addresses content and transformation
provenance in delegated agent workflows, while Inference Chain addresses
computational provenance. draft-liu-oauth-chain-delegation
<https://datatracker.ietf.org/doc/draft-liu-oauth-chain-delegation/>
addresses delegation records with attestation, optional policy and
authorization evidence, while Karl’s draft-mcguinness-oauth-actor-profile
<https://datatracker.ietf.org/doc/draft-mcguinness-oauth-actor-profile/>
and related work address actor representation and evidence layering. The
Liu and McGuinness work could compose with Actor Chain’s Verified Full
Disclosure profile, while Actor Chain remains focused on AS-mediated
actor-path progression, disclosure, and continuity across successive RFC
8693 exchanges.
_______________________________________________
OAuth mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to