DdevOct 16 · 8 min readapexe: Run Your Tools. Enforce Your Rules.Your team already has tools that work: command-line utilities, internal CLIs, operational scripts, and HTTP APIs. They encode years of decisions about how your systems should behave. As AI agents begiblog.aiperceivable.com
PPowerduckOct 8 · 8 min readA2A streaming and the task lifecycle: SSE progress updates, artifacts, and push notificationsWhen an agent takes thirty seconds or five minutes to finish, a plain request/response call is the wrong shape: the connection idles through proxies that time out, the user sees no progress, and a retblogs.powerduck.com
PPowerduckOct 6 · 6 min readBridge an existing HTTP API to A2A agents: JSON-RPC, REST, and gRPC from one handlerMost teams that need an A2A endpoint already own the hard part: the business logic. They have an HTTP service that takes a request, does the work, and returns a result. What they lack is the agent-facblogs.powerduck.com
PPowerduckOct 6 · 7 min readThe A2A Agent Card explained: discovery, supported interfaces, and signature verificationBefore two agents can collaborate, the caller has to answer three questions without a private introduction: where the agent is, what it can do, and how to talk to it. In the Agent2Agent protocol the ablogs.powerduck.com
PPowerduckOct 6 · 7 min readA2A vs MCP in 2026: when agents should talk to agents instead of calling toolsTwo protocols show up in almost every 2026 agent design conversation: MCP and A2A. They are often presented as rivals, which is the first mistake. MCP and Agent2Agent solve different problems at diffeblogs.powerduck.com
MMohammad Hatif OsmaniAug 31 · 17 min readBuilding Corporate, part 4: the final 72 hoursPart 4 of a four-part series on building Corporate, and the last one. Part 1, part 2, and part 3 covered the shape of the system, the department/verification patterns, and the capability expansion witmhosm.hashnode.dev
NNeural CoreTechAug 24 · 2 min readMCP vs A2A in 2026: The Protocol Stack Behind Interoperable AI AgentsThe AI industry is beginning to move beyond the "one model, one agent, one application" architecture. The next stage is more distributed: One agent reasons.Another performs a specialized task.A third 1neuralcoretech.hashnode.dev
PPrashant KoiralaAug 17 · 30 min readAgent-to-agent protocols are becoming a distributed systems layerAI agents are starting to look less like features inside one application and more like services on a network. One agent searches policy documents. Another checks inventory. Another negotiates delivery1blog.prashantkoirala.info.np
KKartik MehtaAug 10 · 15 min readSubagents vs A2A: One Is a Pattern, One Is a ProtocolI get asked to compare these two often enough that it is worth writing down properly. The short answer is that subagents and A2A are not competing, and treating them as alternatives leads people to bu2writer.mrmehta.in
MManu ShuklaAug 7 · 17 min readMCP vs A2A in 2026: 5 tests for whether you need agent-to-agent protocolMCP vs A2A in 2026: 5 tests for whether you need agent-to-agent protocol Summary. The two protocols solve different problems and most teams need only one. The Model Context Protocol connects an agent 11Kecorpit.hashnode.dev
