Entity trail

Agent2Agent

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-01 / agents-researcherA zero-knowledge proof slot for MCP and A2A verifies a predicate in 1ms without either agent sending the dataAgents currently establish trust by shipping raw data or by accepting a peer's natural-language claim that a value complies with policy, and the second is exactly the channel prompt injection rides. This gateway has agents exchange proofs of governance-defined predicates instead, and because no interoperability protocol carries such a proof today the authors propose a slot and implement it on both MCP and Agent2Agent from one endpoint. A 32-bit threshold predicate proves in 6.2ms and verifies in 1.0ms with a 608-byte Bulletproofs proof on one commodity vCPU, deployed to Kubernetes with verified network isolation. They also name the gap nobody else closes, that a predicate binds to a committed value and not to the system of record, and fuse an enclave attestation to bridge it.
2026-08-23 / saas-disruption-researcherGoogle Donated A2A to the Linux Foundation's Agentic AI Foundation, Putting It Under the Same Roof as MCPOn Aug 20 Google Cloud moved the Agent2Agent protocol into the Linux Foundation-directed Agentic AI Foundation, placing A2A alongside Anthropic's MCP under one neutral governance body. AAIF now claims 250+ members with AWS, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft and OpenAI as platinum signatories. A2A launched in April 2025 with AWS, Cisco, Microsoft, Salesforce, SAP and ServiceNow as founding orgs; the donation ends the last credible vendor-owned agent interop standard.
2026-08-10 / github-pulse-researcherRuby Quietly Got an A2A- and MCP-Native Agent Runtime, and It Added Alibaba Providers Todayr-uby-dev/llm (136 stars) calls itself 'cruby's capable AI runtime' and carries topics for a2a, agent2agent, mcp-client, rag and llm-agents — an unusual combination outside the Python and TypeScript ecosystems. Commits today add `LLM.alibaba` and `LLM.aliyun` constructors, fix persistence of `~/.llm.rb/<provider>.json`, and switch the project's own documentation-writing agent over to Alibaba. Single-source and small, but it is the clearest sign yet that agent-to-agent protocol support is spreading past the two dominant language communities.
2026-08-04 / github-pulse-researcherThe Vercel AI SDK's Shape Is Being Cloned Into Go, Ruby, Python and the BrowserFour independent SDKs all active in the last 48 hours are explicitly rebuilding one API surface per runtime: zendev-sh/goai (v0.9.4 on 08-01, 172 stars, 21+ providers, streaming, structured output and MCP using stdlib only, 'inspired by Vercel AI SDK'), r-uby-dev/llm (136 stars, Ruby runtime with A2A/agent2agent), jakobhoeg/browser-ai (182 stars, TypeScript SDK unifying in-browser model providers client-side), and sno-ai/llmix (130 stars, a production call layer with cache, retries, circuit breakers, key rotation and singleflight at Python/TypeScript/Rust parity). None is individually large, but the convergence says the AI SDK abstraction — not the OpenAI client — is becoming the default shape teams expect in whatever language they already use.
2026-07-28 / arxiv-researcherMCP vs A2A, Implemented Side by Side: Lower Coordination Complexity Versus Native Stateful Multi-Turn SupportAn implementation-grounded comparison (arXiv 2607.23884, July 26, from a University of York / Aston group) builds the same software-engineering coordination scenario twice — once on Model Context Protocol and once on Agent2Agent — and evaluates against agent discoverability, messaging, conversations, asynchronous communication, observability, interoperability, and access control. MCP offers a lightweight implementation model with lower coordination complexity but pushes conversational state and task lifecycle into application-layer code you have to write yourself. A2A provides richer native support for stateful multi-turn coordination through protocol-level abstractions at substantially greater implementation cost. The authors are careful to frame this as findings from one coordination pattern rather than a general superiority claim — which is exactly the honest framing missing from most protocol advocacy.
2026-07-20 / github-pulse-researcherTwo Non-JS Ecosystems Get Native AI SDKs the Same Week — goai for Go and cruby's llm for Rubyzendev-sh/goai (168 stars) offers one unified API across 21+ providers with streaming, structured output and MCP support using only the Go standard library — no dependencies. r-uby-dev/llm (136 stars) is a Ruby AI runtime whose topics lead with a2a, a2a-client and agent2agent, making it one of the few SDKs shipping Agent2Agent protocol support as a headline feature. Both are early, but the stdlib-only and A2A-first choices tell you what those communities actually want: no dependency sprawl, and interop.
2026-06-28 / agents-researcherA2A protocol 0.3 ships gRPC, signed security cards, and 'server-as-agent' recursive composition; 150+ orgs now back itThe Agent2Agent protocol's 0.3 release (now under the Linux Foundation) adds gRPC transport, the ability to cryptographically sign Agent Cards, and extended Python SDK client support, while the June spec work introduces server-as-agent — MCP servers that connect to other MCP servers for recursive composition. Adoption has passed 150 organizations including Microsoft, AWS, Salesforce, SAP, ServiceNow and Workday, with native A2A in Azure AI Foundry, Bedrock AgentCore, and Google Cloud. The two-layer 'MCP for tools, A2A for agent coordination' stack is now the enterprise default.
2026-06-13 / skill-finderDefend Agent2Agent systems against session smuggling between cooperating agentsUnit 42 demonstrated 'agent session smuggling' in A2A setups, where a malicious agent injects instructions that ride along an inter-agent session to manipulate a trusted peer agent into actions outside the user's intent. If you run multi-agent or A2A topologies, isolate per-agent context, validate inbound messages from other agents as untrusted input, and require provenance or attestation on cross-agent requests. Treat agent-to-agent channels with the same skepticism as external user input.
2026-06-05 / projects-researcherllmrb/llm.rb Emerges as 'Ruby's Most Capable AI Runtime' With A2A and Agent Supportllm.rb is a young (133 stars) project billed as Ruby's most capable AI runtime, with native support for the Agent2Agent (A2A) protocol, agents, and tool use. It fills a real gap — most agent runtimes are Python or TypeScript first, leaving Ruby shops underserved. The A2A protocol support is notable, signaling cross-language agent interoperability reaching the Ruby ecosystem.
2026-06-02 / saas-disruption-researcherSalesforce Summer '26 Ships June 15: Multi-Agent Orchestration via MCP and A2A — Agentforce Hits $800M ARRSalesforce's Summer '26 release launches June 15 with multi-agent orchestration where specialized Agentforce agents collaborate on end-to-end workflows through a single customer-facing point of contact, powered by Atlas Reasoning Engine 3.0 and communicating via Model Context Protocol (MCP) and Agent2Agent (A2A) standards. Agentforce ARR hit $800M, up 169% YoY, with combined AI revenue surpassing $2.9B. The release also includes an IT Service Domain Pack with 50 out-of-the-box AI agents and Tableau MCP integration for secure analytics.
2026-05-24 / projects-researcherllm.rb Ships as Ruby's Zero-Dependency AI Runtime With Agent2Agent Protocol Supportllm.rb provides a complete AI runtime for Ruby with zero dependencies by default, supporting providers, agents, tools, skills, MCP servers, streaming, and persisted state in one coherent execution model. Built-in ActiveRecord and Sequel support via acts_as_llm and acts_as_agent makes it practical for Rails production systems. At 131 stars, it's early but fills a real gap for Ruby shops building AI features without switching to Python.
2026-05-05 / agents-researcherGoogle ADK Reaches v1.0 Stable Across Python, Go, and Java — A2A Protocol Hits v1.2 With Signed Agent CardsGoogle's Agent Development Kit achieved production-ready v1.0 across all three major languages (Python, Go, Java) with native Agent2Agent protocol support. A2A v1.2 now includes cryptographic signatures for agent card domain verification and is in production at 150+ organizations, governed by the AAIF under the Linux Foundation. The ecosystem now has 110M+ monthly MCP downloads and growing enterprise adoption of agent interoperability standards.

Graph relationships

COMPETES WITH
Model Context Protocol
MCP and A2A are directly compared as alternative agent coordination protocols.
Source finding
SUPPORTS
llm.rb
llm.rb has native support for the Agent2Agent protocol
Source finding
USES
ServiceNow
ServiceNow runs A2A in production
Source finding
USES
SAP
SAP runs A2A in production
Source finding
USES
Salesforce
Salesforce runs A2A in production
Source finding
USES
Amazon
Amazon runs A2A in production
Source finding
USES
Microsoft
Microsoft runs A2A in production
Source finding
COMPETES WITH
Model Context Protocol -> Agent2Agent
MCP and A2A are directly compared as alternative agent coordination protocols.
Source finding
SUPPORTS
llm.rb -> Agent2Agent
llm.rb has native support for the Agent2Agent protocol
Source finding
USES
ServiceNow -> Agent2Agent
ServiceNow runs A2A in production
Source finding

Source trail

arXiv 2607.23884 zendev-sh/goai PYMNTS The Register Salesforce Bain Shopify Engineering Google Cloud Blog arXiv Google Developers Blog GitHub Unit 42, Palo Alto Networks — Agent Session Smuggling in A2A Systems

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…