Generate traceable test suites from project specifications

BMCLI keeps test intent, commands, and evidence in one readable representation: the test script is documentation, and the test case is an inspectable engineering asset.

Version requirements, communication matrices, DBC, CDD, diagnostic surveys, and target variants. Resolve the bench with bmcli bench show --format=json, then inspect database resources through the current database load/list/info/validate Help contracts.

AI can extract signal ranges, timing, diagnostic services, preconditions, and expected responses. Each generated case retains the requirement ID, input combination, BMCLI command, expected result, evidence path, and cleanup action.

BMCLI preserves the high-level meaning of “send, observe, diagnose, and expect” instead of hiding it inside a generic framework. Reviewers can inspect intent without understanding a low-level API, while JSON/JSONL results and logs link back to the same case. This supports the bidirectional traceability expected by processes such as ASPICE.