Add a step to a test case scenario and return {createdStepId, steps_count}.

Exactly one of body (step text), attachment_id, shared_step_id is required. parent_step_id makes the new step a child of an existing step.

Record a test result in a launch and return {ok, id, name, launchId}.

status must be one of failed, broken, passed, skipped, unknown, in_progress; any other value fails here, without a request. duration, start and stop are milliseconds (int64). test_case_id links the result to a test case. JSON-string arguments: tags_json [{"id": 3, "name": "smoke"}] links_json [{"name": "JIRA", "type": "issue", "url": "https://jira/browse/ABC-1"}] custom_fields_json [{"customField": {"id": 12, "name": "Severity"}, "name": "high"}]

Close a launch; the API answers 204. Results are read-only afterwards.

Create a launch and return {ok, id, name, projectId}.

tags_json: [{"id": 3, "name": "regress"}]; links_json: [{"name": "JIRA", "type": "issue", "url": "https://jira/browse/ABC-1"}].

Create a test case and return {ok, id, name, projectId}.

name is required, project_id defaults to TESTOPS_PROJECT_ID. JSON-string arguments: tags_json [{"id": 3, "name": "smoke"}] (the API requires the tag id) links_json [{"name": "JIRA", "type": "issue", "url": "https://jira/browse/ABC-1"}] members_json [{"id": 7, "name": "ivanov", "role": {"id": 1, "name": "MEMBER"}}] (role is an object RoleDto, not a plain string) custom_fields_json [{"customField": {"id": 12, "name": "Severity"}, "name": "high"}] (the API requires both id and name of the custom field) scenario_json {"steps": [{"type": "body", "body": "Open the form"}]} Ids for status_id, workflow_id and test_layer_id come from list_statuses and list_reference. More steps are added with add_test_case_step.

Create a test plan and return {ok, id, name, projectId}.

base_rql is the AQL selection of the plan (e.g. tag.name = "smoke"). tree_selection_json: {"leafsInclude": [1001], "inverted": false} - the plan takes TreeSelectionDto, which has no filterId (that one belongs to running cases).

Delete a launch together with all its results; the API answers 204.

Delete a test case; the API answers 202 and moves it to the trash.

Without force a case that has run results may stay; force=true deletes it anyway, together with its scenario. Restoring a case is not exposed as a tool.

Delete a test plan; the API answers 204. The test cases themselves stay.

Get one launch: id, name, projectId, closed, createdDate, autoclose, releaseId, tags, links.

Status counts of a launch as {passed: n, failed: n, broken: n, skipped: n, ...}.

Cheaper than listing the results: one request instead of paging through them.

Get one project by id: id, name, abbr, isPublic, isVersioned, launchCount, automationPercent.

Get one test case by id: texts, status, workflow, testLayer, tags, links.

The answer does NOT contain the scenario/steps (TestCaseDto has none) and says so with "steps_not_included": true - read them with get_test_case_steps.

Which launches ran a test case, and with what result.

Answers "where was this case run": each row is {testResultId, launchId, launchName, status, start, stop, duration, message, testedBy}. launch_id narrows to one launch, search narrows by result text. project_id defaults to TESTOPS_PROJECT_ID. For the full result of one row use get_test_result(testResultId).

Read the scenario (steps) of a test case.

as_tree=true (default) returns nested steps: {id, body, expectedResult, attachmentId, sharedStepId, children}. as_tree=false returns the raw scenarioSteps map by id. A shared step appears as a node with sharedStepId only - its own steps are not inlined. Attachment contents are not inlined either.

Get one test plan: id, name, projectId, baseRql, testCasesCount, treeSelection.

Get one test result: status, duration, message, trace and the text fields.

The *Html duplicates are dropped. For the small table of a whole launch use list_test_results or get_launch_statistic.

List launches of a project: id, name, closed, createdDate, statistic.

statistic holds status counts as the API returns them, e.g. {passed, failed, broken, skipped, unknown} or StatisticDto with inProgress, pending and total; it is absent when the launch shape carries none. The API answers with two different launch shapes; both are read into this summary. project_id defaults to TESTOPS_PROJECT_ID; sort defaults to "created_date,DESC".

List TestOps projects, ordered by name.

search: optional substring of the project name; favorite: only starred projects. page starts at 0, size is capped at 100. Returns {content, totalElements, totalPages, number} with id, name, abbr, isPublic, isVersioned, launchCount, automationPercent per project.

List a reference dictionary by kind: workflow, testlayer, custom_field or tag.

Returns id and name of every entry - they are what create_test_case wants in workflow_id, test_layer_id and custom_fields_json. custom_field is project scoped (project_id defaults to TESTOPS_PROJECT_ID) and accepts query; tag is a flat list without pagination, so query is applied here.

List test statuses (id, name, color) - they give status_id for create_test_case.

workflow_id narrows the answer to one workflow. The endpoint takes no project id: statuses are defined by workflows, not by projects.

List test cases of a project as rows: id, name, status, automated, flaky, testLayer.

Rows carry no description and no steps - read one case with get_test_case and its steps with get_test_case_steps. project_id defaults to TESTOPS_PROJECT_ID. sort: "property,ASC|DESC", e.g. "createdDate,DESC".

List test plans of a project: id, name, projectId, baseRql, testCasesCount.

project_id defaults to TESTOPS_PROJECT_ID; search narrows by name.

List results of a launch: id, name, testCaseId, status, duration, message, trace, start, stop.

status is one of failed, broken, passed, skipped, unknown, in_progress.

Reopen a closed launch so results can be added again; the API answers 204.

Create a launch with one manual result per selected test case.

Select the cases with one of: ids (comma separated or JSON array of tree leaf ids, becomes selection.leafsInclude), leafs_include (explicit JSON array) or filter_id (saved filter); ids and leafs_include are mutually exclusive. The API field is leafsInclude: there is no leafIds and no rql in the selection. assignees: comma separated user logins; tags_json: [{"id": 3, "name": "regress"}]. Returns {ok, launchId, name, projectId}.

Search test cases with AQL (the API calls it rql) and return full cases.

Give either rql or name. name becomes the AQL query name ~ "" on the server, nothing is filtered client-side. Examples (syntax to be verified on the stand): name ~ "login", status.name = "Draft", tag.name = "smoke". Returns {content: [case fields], rql, totalElements, totalPages, number}.

Update fields of a test case; only the arguments you pass are sent.

An empty string means "leave the field alone" unless clear=true, where an empty string clears the field. JSON-string arguments are the same as in create_test_case.

Update fields of a test plan; only the arguments you pass are sent.

An empty string means "leave the field alone" unless clear=true, where an empty string clears the field. tree_selection_json: {"leafsInclude": [1001], "inverted": false}.

Check an AQL query (the API calls it rql) before running search_test_cases.

Returns {ok, valid, count}: count is how many test cases the query selects, so a typo is caught before a real search. project_id defaults to TESTOPS_PROJECT_ID; deleted=true counts removed cases too. rql is required.