Glossary · A2A concepts

Agent capabilities (A2A)

The capabilities object in an A2A Agent Card: flags for streaming, push notifications and an extended card, plus the list of supported extensions.

Agent capabilities are the optional protocol features an A2A agent declares in its Agent Card’s required capabilities object, so a client knows which operations it may call before it calls them.

Fields.

Field Meaning when true
streaming SendStreamingMessage and SubscribeToTask are available
pushNotifications The agent can POST task updates to a client webhook
extendedAgentCard Authenticated clients can fetch a fuller card with GetExtendedAgentCard
extensions A list of AgentExtension entries, each with a uri, a description, a required flag and optional params

A flag that is absent counts as false.

Enforcement. Section 3.3.4 turns these flags into rules. If streaming is false or missing, streaming operations must fail with UnsupportedOperationError (JSON-RPC -32004). If pushNotifications is false or missing, all four push notification config operations must fail with PushNotificationNotSupportedError (-32003). If extendedAgentCard is false or missing, GetExtendedAgentCard must fail with UnsupportedOperationError. An agent that declares the extended card but has none configured returns ExtendedAgentCardNotConfiguredError (-32007). When an extension is marked required and the client does not declare support for it, the agent returns ExtensionSupportRequiredError (-32008). Clients should read the card first.

A live example. Emissar’s public agent card, fetched on 26 September 2026, declares:

{"capabilities": {"streaming": false, "pushNotifications": false, "extendedAgentCard": false}}

A client following that card polls with GetTask and does not open streams or register webhooks.

Version changes. In v0.3 the extended card flag was a top-level card field, supportsAuthenticatedExtendedCard, and Appendix A.2.2 records its move into capabilities as a breaking change. The v0.3 stateTransitionHistory capability is not part of v1.0.

Neighbouring terms. Capabilities describe protocol features. Skills describe what the agent can do for a caller, and securitySchemes describes how to authenticate.

Sources

  1. A2A protocol definition (a2a.proto): AgentCapabilities, AgentExtension (accessed )
  2. A2A Protocol Specification, section 3.3.4: Capability Validation (accessed )
  3. A2A Protocol Specification, Appendix A.2.2: Extended Agent Card Field Relocated (accessed )
  4. A2A v0.3.0 type definitions (types.ts): AgentCapabilities (accessed )
  5. Emissar public Agent Card (live) (accessed )