Engineering Notes

SAMP Engineering Notes #12

What does SAMP conformance actually mean? Explore role-specific tiers, mandatory requirements, reproducible test matrices, failure injection, and evidence-based interoperability claims.

SAMP event delivery and audit boundaries, showing PUSH notifications, subscriptions, duplicate detection, transport failures, durable recording, and operational accountability.

Engineering Notes

SAMP Engineering Notes #11

Receiving an event is not the same as recording it. Explore SAMP PUSH notifications, subscription ownership, duplicate delivery, recovery limits, and the boundary between events and durable audit records.

SAMP agent profiles as versioned management contracts, showing revisions, authenticated announcement ordering, schema validation, revision-bound queries, and safe recovery.

Engineering Notes

SAMP Engineering Notes #10

How SAMP treats agent profiles as versioned management contracts, with authenticated announcements, schema validation, revision-bound queries, and safe recovery without confusing capabilities with authority.

Technical comparison of MCP, A2A and SAMP illustrating tool and context integration, agent-to-agent collaboration, and AI agent management as distinct complementary responsibilities.

Engineering Notes

SAMP Engineering Notes #09

MCP, A2A and SAMP address different responsibilities in AI systems. Explore the distinction between tool integration, agent collaboration, and AI agent management.

Technical illustration of SAMP AI agent authorization showing authentication, trust, policy evaluation, human approval, exact-action authorization, and policy enforcement before execution.

Engineering Notes

SAMP Engineering Notes #08

Understanding an instruction is not the same as being authorized to execute it. Explore how SAMP separates authentication, trust, exact-action authorization, human authority, and policy enforcement.

Engineering Notes

SAMP Engineering Notes #07

What happens when an AI agent crashes, becomes unreachable, or leaves an operation's outcome uncertain? Explore SAMP failure management, recovery, trust, enforcement, and lessons from AgMS Alpha.

AgMS Alpha experimental AI agent management environment featuring Codex CLI, OpenClaw, Hermes, Agent Zero and GROK Bot, with monitoring, operational events and failure testing.

Engineering Notes

SAMP Engineering Notes #06

From protocol design to real-world experimentation. Discover how AgMS Alpha applies SAMP management concepts to heterogeneous AI agents, monitoring, operational events, and failure testing.

Engineering Notes

SAMP Engineering Notes #05

An authorized AI-agent operation is not necessarily executed, and execution is not necessarily verified. Explore how SAMP combines exact-action authorization, runtime evidence, recorded outcomes, and safe recovery.

SAMP operational management showing independent agent trust, liveness, quarantine enforcement status, and deployment failure domains.

Engineering Notes

SAMP Engineering Notes #04

An AI agent can be reachable but untrusted, trusted but offline, or quarantined without being isolated. Explore how SAMP separates trust, liveness, enforcement, and failure domains.

SAMP Management Plane architecture showing Manager and Agent roles, eight protocol operations, and operational trust and liveness.

Engineering Notes

SAMP Engineering Notes #03

Inside the SAMP Management Plane: how Managers and Agents cooperate through eight protocol operations to provide discovery, monitoring, trust management, controlled execution, and operational accountability.

Engineering Notes

SAMP Engineering Notes #02

From managing routers and switches to managing AI agents. Discover how decades of network management experience inspired SAMP's approach to monitoring, trust, authorization, operational control, and failure handling.

Engineering Notes

The Evolution of SAMP: From -00 to -04

From the first Internet-Draft to SAMP v0.2: five revisions exploring AI agent management, trust, authorization, runtime evidence, and operational resilience.