{
  "schema": "durable-workflow.v2.platform-conformance.runtime-scenarios",
  "version": 1,
  "category": "search_attribute_runtime_contract",
  "suite_schema": "durable-workflow.v2.platform-conformance.suite",
  "suite_version": 38,
  "description": "Public scenario manifest for Temporal-parity search-attribute conformance. Harnesses use these scenario ids and criteria when evaluating published Durable Workflow artifacts.",
  "fixture_type": "published_artifact_runtime_scenario_manifest",
  "result_statuses": [
    "pass",
    "fail",
    "unsupported",
    "not_covered",
    "runner_blocked"
  ],
  "artifact_policy": {
    "published_artifacts_only": true,
    "requires_resolved_versions": true,
    "rejects_placeholder_versions": true,
    "implementation_tests_are_not_sources": true
  },
  "common_result_evidence": [
    "suite_version",
    "scenario_id",
    "status",
    "started_at",
    "finished_at",
    "generated_at",
    "published_artifact_versions",
    "implementation_identity",
    "runtime_matrix",
    "topology",
    "query_verdicts",
    "observed_outputs",
    "linked_findings"
  ],
  "scenarios": [
    {
      "id": "published_artifact_install_only",
      "title": "Published artifact install only",
      "claim_targets": [
        "standalone_server",
        "official_sdk",
        "worker_protocol_implementation",
        "cli_json_client",
        "waterline_contract_surface"
      ],
      "actors": [
        "server",
        "cli",
        "python_sdk_runtime",
        "sdk_php_runtime",
        "waterline"
      ],
      "operations": [
        "install the server image, CLI, Python SDK, PHP SDK runtime, and Waterline from published channels",
        "record the resolved artifact name and version for every actor",
        "run no product code from a local checkout"
      ],
      "pass_criteria": [
        "the result document records every resolved artifact version",
        "all actors under test came from published install channels",
        "local product source checkouts are not used as artifacts under test"
      ],
      "unsupported_when": [
        "the implementation does not claim the actor or target under test"
      ]
    },
    {
      "id": "schema_definition_and_reserved_name_refusal",
      "title": "Schema definition and reserved-name refusal",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "define string, int, double, bool, datetime, keyword, and keyword-list search-attribute keys in one namespace",
        "read the schema back through the server, CLI, and at least one SDK client path",
        "attempt to define reserved or system-prefixed names including wf_id and __internal"
      ],
      "pass_criteria": [
        "all documented search-attribute types are durable and readable through public surfaces",
        "reserved names are rejected with typed user-facing errors",
        "schema evidence is scoped to the namespace under test"
      ]
    },
    {
      "id": "python_worker_start_and_upsert_visibility",
      "title": "Python worker start and upsert visibility",
      "claim_targets": [
        "standalone_server",
        "official_sdk",
        "worker_protocol_implementation",
        "cli_json_client"
      ],
      "actors": [
        "server",
        "python_worker",
        "python_sdk_client",
        "cli_client"
      ],
      "operations": [
        "start Python-worker runs with typed search attributes at workflow start",
        "upsert at least one attribute from workflow code while a run is executing",
        "query the start and upserted values through public visibility paths"
      ],
      "pass_criteria": [
        "start attributes and upserted attributes are both visible by query",
        "the merged value set matches the workflow source data exactly",
        "CLI and Python SDK readers identify the same workflow runs"
      ]
    },
    {
      "id": "php_worker_start_and_upsert_visibility",
      "title": "PHP worker start and upsert visibility",
      "claim_targets": [
        "standalone_server",
        "official_sdk",
        "worker_protocol_implementation",
        "cli_json_client"
      ],
      "actors": [
        "server",
        "php_worker",
        "php_sdk_client",
        "cli_client"
      ],
      "operations": [
        "start PHP-worker runs with typed search attributes at workflow start",
        "upsert at least one attribute from PHP workflow code while a run is executing",
        "query the start and upserted values through public visibility paths"
      ],
      "pass_criteria": [
        "start attributes and upserted attributes are both visible by query",
        "the merged value set matches the workflow source data exactly",
        "CLI and PHP SDK readers identify the same workflow runs"
      ]
    },
    {
      "id": "cli_query_and_error_surface",
      "title": "CLI query and error surface",
      "claim_targets": [
        "cli_json_client",
        "standalone_server"
      ],
      "actors": [
        "server",
        "cli_client"
      ],
      "operations": [
        "create, list, and delete custom search-attribute definitions through the CLI",
        "list workflows with equality, range, boolean, keyword-list, OR, and NOT query expressions",
        "submit a malformed or type-invalid query and capture the CLI error"
      ],
      "pass_criteria": [
        "CLI JSON and human-readable output expose successful schema operations",
        "workflow list query results match the source dataset counts exactly",
        "query syntax or type errors surface as typed operator-facing errors"
      ]
    },
    {
      "id": "waterline_operator_visibility",
      "title": "Waterline operator visibility",
      "claim_targets": [
        "waterline_contract_surface",
        "standalone_server"
      ],
      "actors": [
        "server",
        "waterline"
      ],
      "operations": [
        "filter the workflow list by a search-attribute key and value",
        "inspect a selected run detail view containing typed search attributes",
        "save and reload a Waterline view with a search-attribute filter"
      ],
      "pass_criteria": [
        "Waterline list filters return the same runs as the server query surface",
        "selected run detail shows the merged search-attribute map",
        "saved filter state round-trips without dropping the search-attribute predicate"
      ]
    },
    {
      "id": "python_to_php_codec_round_trip",
      "title": "Python to PHP codec round trip",
      "claim_targets": [
        "official_sdk",
        "worker_protocol_implementation",
        "cli_json_client"
      ],
      "actors": [
        "server",
        "python_worker",
        "php_sdk_client",
        "cli_client"
      ],
      "operations": [
        "write all supported search-attribute value types from a Python workflow or client",
        "read and verify those values from a PHP-facing client path",
        "mirror the verification through CLI workflow list or describe output"
      ],
      "pass_criteria": [
        "PHP-facing readers observe the same scalar, datetime, and keyword-list values written by Python",
        "no value changes type, precision, timezone, or list ordering semantics during round trip",
        "CLI evidence agrees with the PHP-facing reader"
      ]
    },
    {
      "id": "php_to_python_codec_round_trip",
      "title": "PHP to Python codec round trip",
      "claim_targets": [
        "official_sdk",
        "worker_protocol_implementation",
        "cli_json_client"
      ],
      "actors": [
        "server",
        "php_worker",
        "python_sdk_client",
        "cli_client"
      ],
      "operations": [
        "write all supported search-attribute value types from a PHP workflow or client",
        "read and verify those values from a Python client path",
        "mirror the verification through CLI workflow list or describe output"
      ],
      "pass_criteria": [
        "Python readers observe the same scalar, datetime, and keyword-list values written by PHP",
        "no value changes type, precision, timezone, or list ordering semantics during round trip",
        "CLI evidence agrees with the Python reader"
      ]
    },
    {
      "id": "equality_range_bool_query_behavior",
      "title": "Equality, range, and bool query behavior",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "seed a known workflow dataset with string, int, double, bool, datetime, keyword, and keyword-list values",
        "query exact string or keyword matches",
        "query numeric and datetime ranges plus boolean predicates"
      ],
      "pass_criteria": [
        "every query reports expected and actual counts",
        "actual counts equal expected counts for the source dataset",
        "returned run ids are a subset of the matching source records with no over-return"
      ]
    },
    {
      "id": "or_not_query_grammar",
      "title": "OR and NOT query grammar",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "query disjunctions with OR across two attribute values",
        "query conjunctions that include NOT against a boolean or keyword predicate",
        "verify operator precedence using a source dataset with distinguishable matches"
      ],
      "pass_criteria": [
        "OR expressions include every matching source run and no non-matching runs",
        "NOT expressions exclude the negated predicate without filtering unrelated matches",
        "server, CLI, and SDK query paths agree on the same run set"
      ]
    },
    {
      "id": "keyword_list_membership",
      "title": "Keyword-list membership",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "set keyword-list attributes with the same values in different orders",
        "query membership for one keyword-list value",
        "inspect returned values through describe or SDK read paths"
      ],
      "pass_criteria": [
        "membership queries match all runs containing the requested value",
        "list ordering does not affect membership matching",
        "keyword-list values remain arrays in read responses"
      ]
    },
    {
      "id": "type_safety_wrong_literal",
      "title": "Wrong literal type safety",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "submit a query comparing an integer search attribute to a string literal",
        "submit a query comparing a bool search attribute to a non-bool literal",
        "capture server, CLI, and SDK error envelopes"
      ],
      "pass_criteria": [
        "wrong-type literals fail with typed errors",
        "the query is not silently coerced",
        "the query does not return an empty success result that hides the type error"
      ]
    },
    {
      "id": "undefined_key_rejection",
      "title": "Undefined key rejection",
      "claim_targets": [
        "standalone_server",
        "official_sdk",
        "worker_protocol_implementation"
      ],
      "actors": [
        "server",
        "worker",
        "sdk_client"
      ],
      "operations": [
        "attempt to start a workflow with an undefined search-attribute key",
        "attempt to upsert an undefined key from workflow code",
        "inspect the run state and history after rejection"
      ],
      "pass_criteria": [
        "undefined keys are rejected with typed errors",
        "the workflow does not advance with bad search-attribute state",
        "history and visibility do not persist the undefined key"
      ]
    },
    {
      "id": "indexing_latency_distribution",
      "title": "Indexing latency distribution",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "worker",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "set or upsert a search attribute and poll the query path until it is visible",
        "repeat the measurement at least twenty times",
        "record min, p50, p95, max, sample count, and the documented latency bound"
      ],
      "pass_criteria": [
        "every sample becomes queryable within the documented bound",
        "the result records the full latency distribution",
        "query results after visibility match the source run exactly"
      ]
    },
    {
      "id": "load_and_bounded_latency",
      "title": "Load and bounded latency",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "worker",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "seed at least one thousand workflow runs with varied search attributes",
        "run representative equality, range, OR, NOT, and keyword-list queries",
        "record p50, p95, and max latency under the load profile"
      ],
      "pass_criteria": [
        "all queries return exact source-dataset matches under load",
        "latency remains within the documented or result-declared bounded profile",
        "the result identifies workflow count and latency distribution fields"
      ]
    },
    {
      "id": "namespace_isolation",
      "title": "Namespace isolation",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk",
        "waterline_contract_surface"
      ],
      "actors": [
        "server",
        "cli_client",
        "sdk_client",
        "waterline"
      ],
      "operations": [
        "define the same search-attribute name in two namespaces",
        "start distinguishable workflows with that attribute in each namespace",
        "query definitions and values from each namespace and through Waterline when claimed"
      ],
      "pass_criteria": [
        "schema reads are scoped to the selected namespace",
        "value queries never return runs from another namespace",
        "Waterline and API visibility agree on namespace-scoped results"
      ]
    },
    {
      "id": "query_injection_hardening",
      "title": "Query injection hardening",
      "claim_targets": [
        "standalone_server",
        "cli_json_client",
        "official_sdk"
      ],
      "actors": [
        "server",
        "cli_client",
        "sdk_client"
      ],
      "operations": [
        "submit a query containing OR 1=1 outside the documented grammar",
        "submit a query containing an embedded SQL comment marker",
        "submit a query containing shell metacharacters after a valid predicate"
      ],
      "pass_criteria": [
        "all injection probes are rejected with typed syntax errors",
        "no rejected query partially executes or returns a result set",
        "error evidence is consistent across server, CLI, and SDK surfaces"
      ]
    }
  ]
}