Conformance blocker: complete Python SDK parity coverage beyond smoke

Author: durable-workflow-ops

CHANGELOG.md

@@ -7,6 +7,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [Unreleased]
 
 ### Added
+- `durable-workflow-python-conformance --compose` now accepts runner-native
+  host evidence aliases such as `scenarioEvidence`, `capabilityResults`,
+  `officialCli`, `firstUserFlow`, `traces`, and `languageNeutralityAudit`,
+  so full published-artifact runs are not marked uncovered just because the
+  runner records observations under natural collection names.
 - `durable-workflow-python-conformance` now exposes a host-evidence
   composition contract and `--compose` mode so published-artifact runners can
   turn raw observations into a complete Python parity result document before

CONFORMANCE.md

@@ -68,6 +68,10 @@ The composer materializes one row for every required scenario and capability.
 Rows with no recorded evidence are emitted as `not_covered`, preserving a
 non-passing result with focused gate failures instead of letting smoke-only
 evidence look complete.
+Runner-native aliases such as `scenarioEvidence`, `capabilityResults`,
+`officialCli`, `firstUserFlow`, `traces`, and `languageNeutralityAudit` are
+accepted as input to the composer; they are normalized into the canonical
+result fields before evaluation.
 
 The result gate requires one result for every scenario below, all marked
 `pass`, before the Python SDK conformance property can pass: