SAP Testing Tools

These are the tools Copilot calls while working through the workflow. You never invoke them yourself — they're listed here so you recognise the names when Copilot mentions one, and know what it just did.

They appear only when SAP Testing is enabled, and they sit alongside the regular ABAP FS tools that Copilot uses for reading source and querying tables.

Tool	What it does
get_test_folder	Returns your configured test folder. Copilot calls this before touching anything, and warns you if the folder isn't open in your workspace
get_sap_webgui_url	Builds a ready-to-open SAP WebGUI URL for a connection, already signed in, optionally landing on a specific transaction
build_test_index	Validates every test case file and rebuilds the case list — _index.md and the printable _index.docx. Refuses to run until the test plan reviewer has passed the plan
build_test_index_docx	Regenerates just the printable _index.docx, for when only the notes changed
split_test_cases	Splits one bulk-authored file into individual test case files, validating each before writing
verify_test_data_usage	Cross-checks a script against its case and data spec: data keys must match, and every table in se16nTables must have matching SE16N proof
check_test_data	Pre-flight check: resolves every case's data for a program on a given system and reports what's missing — before you waste a test run finding out
playwright_test	Replaces terminal npx playwright test. Runs exact tcIds or a whole program; independent cases may run across up to 5 workers after one login, while dependent/shared-state cases stay in ordered sequential batches. Stops after 3 failures by default (maxFailures cap 10) and rejects incomplete SE16N proof
build_evidence_report	Builds the aggregated Word evidence report for a program and system from all the run results
analyze_anst_enhancements	Classifies an ANST export and writes a work list beside it

Two tools that push back

build_test_index and playwright_test are deliberately gated: Copilot has to certify that the required review or readiness check actually happened before either will run. If Copilot tries to skip ahead, the tool rejects the call.

This is why you'll sometimes see Copilot go back and do something it seemed to have finished — the tool told it to. Details in the Technical Reference.

If Copilot says a tool is missing

VS Code doesn't always surface every tool to the model straight away, and playwright_test is the one that most often goes missing. Copilot is instructed to search for a tool by name before giving up, and to tell you rather than improvise — so if it reports a missing tool, that's a genuine report, not a mistake.

Starting a new chat usually clears it. What Copilot must never do is substitute a terminal command; there is no npx playwright test route here.