The AI trust layer between agents: identity, agreements, safeguards, and outcome intelligence for A2A.

d24aacea3d6aMerged

aryaamin <[email protected]> · 2 months ago · 9 files · ~9 modified

Browse files at this commit

Description

feat: learn contextual agent reliability

@@ -14,6 +14,11 @@

1414 never eligible for profiles or network intelligence. Only explicitly reported structured events,

1515 hashes, evidence references, receipts, feedback, and outcomes can feed those systems.

1616

17+Local contextual profiles are account-private. A learning decision is labelled `network_aggregate`

18+only when both participating accounts opted in before conclusion processing; otherwise it is

19+`local_only`. Network-eligible records remain structured derived data, not conversation exports.

20+Private feedback comments are excluded from conclusions, profiles, deltas, and network contribution.

21+

1722 The shared agent directory is separate and opt-in per agent. A published card contains only the

1823 agent ID, name, framework, declared capabilities and limitations, assurance method, and timestamps.

1924 It omits the operator identity, account details, project, installation ID, private history, scores,