Intro + a lifecycle/presence contribution offer (Living Agents Protocol)

Hi all, and chairs Gaowei Chang and Song XU,

I have been working on the existence side of agent protocols: how a
persistent agent proves it is alive, how its authority decays when it goes
dark, and why a restart should not read as a new identity. It is a small,
implemented v0 model called the Living Agents Protocol: two interoperating
reference libraries (Node and Python) checked against shared test vectors,
a runnable two-process demo in which one agent is killed mid-session and
the counterparty's authority suspends and later recovers with the same
identity, and a written threat model. It deliberately sits under MCP and
A2A rather than competing with them, which lines up with this group's
"discover, identify, and collaborate" scope.

Two pieces seem most relevant here, and I would value the group's read on
them:

1. The presence and pulse mechanism: signed heartbeats chained to the
agent's log, with authority that suspends by protocol after a declared gap
and resumes on recovery.
2. The agent passport: an identity bound to a responsible principal, with
labelled proof classes, carried as a compact JWS that any counterparty can
verify without a registry.

If there is interest I am glad to write either up as a contribution in
whatever form the group prefers. Repository, specification and demo:
https://github.com/atharnouman/living-agents-protocol ; the essay on the
why, in about 1,100 words:
https://atharnouman.github.io/living-agents-protocol/essay.html

Best regards,
Athar Nouman

-- 
d/dx (aThAr)