Learn Requirement

Learn About a Requirement

Follow these steps to help the user understand a TCK requirement in depth.

Step 1: Identify the requirement

Ask the user for the requirement ID (e.g., CORE-SEND-001, GRPC-ERR-001, JSONRPC-SSE-001, HTTP_JSON-REST-001).

If the user describes a concept instead of an ID, search the requirement files to find matching requirements:

tck/requirements/core_operations.py    — CORE-*
tck/requirements/binding_grpc.py       — GRPC-*
tck/requirements/binding_jsonrpc.py    — JSONRPC-*
tck/requirements/binding_http_json.py  — HTTP_JSON-*
tck/requirements/data_model.py         — data model requirements
tck/requirements/streaming.py          — streaming requirements
tck/requirements/push_notifications.py — push notification requirements
tck/requirements/agent_card.py         — agent card requirements
tck/requirements/auth.py               — authentication requirements
tck/requirements/versioning.py         — versioning requirements
tck/requirements/interop.py            — interoperability requirements

Step 2: Read the requirement definition

Find and read the RequirementSpec entry for the requirement ID in the appropriate file under tck/requirements/.

Present the following fields to the user:

• ID — the requirement identifier
• Section — the spec section number
• Title — short description
• Level — MUST, SHOULD, or MAY
• Description — what the requirement says
• Operation — which A2A operation it applies to
• Expected behavior — what a compliant implementation should do
• Spec URL — convert the local spec_url to a GitHub link: read specification/version.json for sourceUrl, replace /tree/ with /blob/, replace trailing /specification with /docs, then append the filename and anchor from spec_url
• Tags — categorization tags
• Sample input — example request payload (if provided)

Step 3: Find the specification text

Read specification/specification.md and search for the section referenced by the requirement's section field. Present the relevant specification text so the user can see the normative language (MUST/SHOULD/MAY) in context.

Step 4: Find related tests

Many requirements do not have a dedicated test function that references their ID by name. Instead, they are tested through the parametrized runner in tests/compatibility/core_operations/test_requirements.py. This runner:

1. Collects all registered RequirementSpec entries (grouped by level: MUST, SHOULD, MAY)
2. Parametrizes each across all transports (grpc, jsonrpc, http_json)
3. Uses execute_operation(client, requirement) to send the requirement's sample_input via the appropriate operation
4. Validates the response (success, schema conformance, or expected error)

Requirements tagged multi-operation or without an operation are excluded from the parametrized runner and are instead covered by dedicated test modules.

To find how a requirement is tested:

1. First, search for the requirement ID in the test files:

   • tests/compatibility/core_operations/ — cross-transport tests
   • tests/compatibility/grpc/ — gRPC-specific tests
   • tests/compatibility/jsonrpc/ — JSON-RPC-specific tests
   • tests/compatibility/http_json/ — HTTP+JSON-specific tests

2. If no dedicated test is found, check whether the requirement has an operation and is not tagged multi-operation. If so, it is tested by the parametrized runner in test_requirements.py.

Read the test code and explain:

• What the test does step by step
• What request it sends to the SUT
• What response it expects
• What conditions cause the test to pass or fail

Step 5: Find related validators

Search tck/validators/ for any validation logic related to this requirement's operation or error codes.

Step 6: Summarize

Present a concise summary to the user:

1. What it requires — plain-language explanation of the requirement
2. Why it matters — the purpose in the context of the A2A protocol
3. How it's tested — what the TCK does to verify compatibility
4. How to pass — what an SUT implementation needs to do to satisfy this requirement