Skip to content

draft: state the isolation premise; define Sandbox; name the identity candidates - #23

Merged
rappdw merged 1 commit into
mainfrom
review-round-2
Oct 21, 2026
Merged

rappdw merged 1 commit into
mainfrom
review-round-2

Conversation

@rappdw

@rappdw rappdw commented Oct 21, 2026

Copy link
Copy Markdown
Member
  • Introduction, "The approach": a paragraph stating the premise — the Agent
    runs in a sandbox whose only path to its correspondence is this protocol;
    isolation makes the Agent operable as well as safe; the document specifies
    what crosses the boundary, not the sandbox.
  • Terminology: Sandbox, the isolation boundary the Namespace is mounted
    into; the mechanism is presumed, not specified. The term was used eight
    times without a definition.
  • "Why This Work Is Brought to the IETF": the identity open question now
    names what each candidate would replace (SPIFFE the configured keys;
    JWS/COSE the statement encoding) and why DKIM/ARC do not fit.

No normative sentence changed; gates unchanged.

🤖 Generated with Claude Code

… candidates

- Introduction, "The approach": a paragraph stating the premise — the Agent
  runs in a sandbox whose only path to its correspondence is this protocol;
  isolation makes the Agent operable as well as safe; the document specifies
  what crosses the boundary, not the sandbox.
- Terminology: Sandbox, the isolation boundary the Namespace is mounted
  into; the mechanism is presumed, not specified. The term was used eight
  times without a definition.
- "Why This Work Is Brought to the IETF": the identity open question now
  names what each candidate would replace (SPIFFE the configured keys;
  JWS/COSE the statement encoding) and why DKIM/ARC do not fit.

No normative sentence changed; gates unchanged.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
@rappdw
rappdw merged commit 2d9919f into main Oct 21, 2026
1 check passed
@rappdw
rappdw deleted the review-round-2 branch October 21, 2026 22:58
rappdw added a commit that referenced this pull request Oct 23, 2026
* site: the protocol is AMMP; the -00 is posted; the -01 questions

Rename the explainer site to AMMP, the Agent Mediated Messaging Protocol,
matching draft-rapp-ammp-00. The amap-* repository names, the AMAP-Peer-*
header fields and the validator's JavaScript identifiers keep their names,
and the home page, FAQ and glossary say why.

Replace "not yet submitted" everywhere: the -00 was posted to the IETF
Datatracker on 2026-10-22 as an individual Internet-Draft, adopted by no
working group. Link the Datatracker from the draft links and the footer,
and name IETF 127 as the next meeting.

Carry the draft's own open-questions list onto the IETF page, each linked to
its issue in the draft-rapp-ammp-01 milestone; state the isolation premise
and define Sandbox (from #23) on the how-it-works, implementers and glossary
pages; add the identity-between-runtimes candidates to the identity page;
add two FAQ entries (why not an SMTP/IMAP proxy; why AMMP when the repos are
amap-*).

build_site.py ignores .DS_Store when copying and comparing, so a Finder
dropping no longer fails --check.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

* site: add the Codex connector beside the Claude Code connector

amap-connector-codex is public under Apache 2.0 and pinned in the same
release manifest. Name it wherever the site named only the Claude Code
connector: getting started, the implementers' reference implementations,
the repository lists, the FAQ, the layering page, and the IETF page's count
of reference implementations. The status page's line on later deployments
now says what it was: the Codex connector running in the same fleet with
the reference runtime, delegating to and answering Claude agents.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant