About — Agent Cards Explained
A reference layer for the A2A Agent Card spec, built and maintained by the AgentsBooks team.
Why this site exists
The A2A v1.0 specification is excellent reference material but, like every well-written spec, it expects the reader to already understand the context. Agent Cards are the load-bearing artefact of A2A — the JSON document at the well-known URL — and the operational decisions around them (how to sign, where to publish, when to rotate keys, how to compare to OpenAPI / robots.txt / model cards / MCP descriptors) are scattered across vendor blogs, IETF drafts, regulator publications, and academic papers. This satellite consolidates the operational knowledge into one place so an engineer can read it once and ship correctly.
Who it is for
Three audiences, in declining order of expected fit:
- Agent developers deciding how to expose their A2A agents to outside callers — what to put in the card, how to sign it, where to host it.
- Marketplace operators building or maintaining catalogs of A2A agents — what to consume from the card, how to verify Signed Agent Cards, how to surface them to buyers.
- Security architects reviewing A2A deployments — what threat vectors the card mitigates, what it does not, where the cross-domain delegation gap actually lies.
The framing
"A2A provides the definitive common language for agent interoperability in a world where agents are built using diverse frameworks and by different vendors."
"Agents can advertise their capabilities using an 'Agent Card' in JSON format, allowing the client agent to identify the best agent that can perform a task and leverage A2A to communicate with the remote agent."
Eight primitives
The vocabulary AgentsBooks uses for everything: Identity · Brain · Heart · Memory · Control · Knowledge · Friends · Shares. See Anatomy of a Firm. Agent Cards live in the Friends primitive — agent-to-agent — and reinforce Identity, Heart, and Control.
Built by AgentsBooks
This is one of a small, growing family of authority properties published by AgentsBooks — pillar essays, comparison matrices, ROI calculators, glossaries, and research indexes. They share a vocabulary, an editorial voice, and a single primary CTA: try AgentsBooks free. The pillar essay at agentsbooks.com/blog/a2a-protocol-explained is the canonical home of the topic; this satellite is its long-tail amplifier on the single Agent Card artefact.
Editorial discipline
Every claim is anchored to a canonical primary source — vendor docs, standards bodies, regulators, peer-reviewed papers. We publish a methodology page with the full bibliography and the freshness flag for each source. We do not editorialise on competitors; we cite their published claims. We acknowledge tradeoffs and counter-narratives where they exist (Marcus 2026 on agent reliability is the canonical skeptic anchor). The full sourcing rule lives on the methodology page; the audit trail (sources surveyed, quotables captured, stale claims retired, open questions) lives in the satellite's source repository at raw/research-notes.md.
Refresh discipline
Default cadence is bi-annual full review (per the build-authority-property workflow's pillar-mini-site cadence). Trigger refreshes on: A2A spec revision (v1.1+); IETF AAuth RFC progression; NIST CAISI publication; Microsoft Entra Agent ID major update; new Linux Foundation founding-member announcement; major marketplace launch.