apidog-test-case

API Test Cases

Prerequisite: read ../apidog-cli/SKILL.md first. If the stable entry skill conflicts with this domain skill, use the current CLI help and this skill as the source of truth. For API definitions, follow current CLI help for endpoint, schema, folder, and related commands. For multi-step flows, read ../apidog-test-scenario/SKILL.md.

Always use the current CLI help for exact arguments. When creating or updating API test cases, pay special attention to categoryId display risks, the boundary between test-case and test-scenario, requestBody/processor/assertion/extractor structure, and run-time verification. Agents must validate payloads with cli-schema validate before writing, then use get to confirm the saved structure.

When To Use

• Create automated test cases under a single endpoint.
• Update test steps, assertions, variable extractors, and pre/post processors.
• List test cases under an endpoint.
• Run API test cases by caseId, endpointId, or categoryId.
• Create or maintain test datasets.
• Diagnose issues such as missing test steps in the client, ineffective assertions, or empty extracted variables.

When Not To Use

• Multi-endpoint workflow orchestration or scenario modeling: use apidog-test-scenario.
• Only running existing scenarios, suites, or CI commands: use apidog-test-automation.
• Only editing API definitions, request parameters, or response models: use endpoint/schema commands from current CLI help.
• Only viewing test reports: use test-report from current CLI help; use apidog-test-automation for execution and report boundaries.

Core Concepts

Command Entry Points

Use current CLI help for test-case, test-data, and apidog run --test-case. Use test-case category to obtain categoryId; current test-case category does not support --endpoint. To view cases under an endpoint, use test-case list --endpoint <endpointId>. apidog run --test-case accepts only caseId.

When importing a single-endpoint case into a test scenario, do not handwrite scenario steps in this skill. Switch to apidog-test-scenario and use:

apidog test-scenario import-steps <scenarioId> --project <projectId> --source test-case --endpoint <endpointId> --ids <testCaseId

Standard Create Flow

1. Confirm project and branch.
2. Locate the endpoint with apidog endpoint list/get.
3. Run apidog test-case category --project <projectId> to obtain a valid categoryId; to inspect existing endpoint cases, run apidog test-case list --project <projectId> --endpoint <endpointId>.
4. If a similar case exists, use list and then get one as a template.
5. Fetch and validate the test-case-create schema.
6. Build a complete JSON payload; do not create an empty shell with only name/endpointId.
7. After creation, immediately run test-case get <caseId> and confirm the saved backend structure.
8. Run once with a local JSON report to confirm requestBody, processors, assertions, and scripts work in the runner.
9. If uploading reports, use test-report according to current CLI help; if report details are missing, use apidog-test-automation for local/cloud report boundaries.

categoryId is a key required field for client display, not an ordinary optional category. An invalid categoryId may make a case visible to CLI get/list but invisible or unusable in the client category list. Always obtain a valid ID with test-case category before creating a case.

Standard Update Flow

Before updating, always get the original structure and modify that complete structure. Then validate against test-case-update schema to avoid losing existing steps, assertions, variable extractors, or processors. update is not JSON Patch and does not merge array elements by id.

The test-case-create and test-case-update schemas include processor, assertion, extractor, and enum documentation. When writing processors for the first time, read the schema first; do not guess field names, enum values, or legacy formats. Current test-case-update validates processor data by type; for example, extractor validates variableType/subject/shareScope, assertion validates subject/comparison, and delay requires numeric data.

test-case get may return an empty string for method. This does not mean the client displays the HTTP method incorrectly; the client usually displays it from the bound endpoint. Current test-case-update schema accepts this readback structure. Do not guess and rewrite the method unless the user explicitly asks to change the bound endpoint or request method.

Test Step Display Rules

If the user cares about client display, verify all of the following:

1. Run test-case get after create or update.
2. Confirm steps, assertions, extractors, and processors were saved by the backend.
3. Confirm fields are not empty arrays, empty objects, or written to the wrong level.
4. If CLI returns success but the client does not display the data, switch to apidog-cli-checkup and first confirm project, branch, endpoint, categoryId, and readback structure.

Content And Processor Structure

• requestBody.data must be a string; do not write JSON Body as an object.
• Format multiline JSON bodies, pre-scripts, and post-scripts with \n; this improves client readability without changing execution semantics.
• preProcessors and postProcessors use the flat structure { id, type, data, defaultEnable, enable }; do not use legacy nested { type, config }.
• Prefer stable id values for processors, especially assertion, extractor, and customScript; missing IDs may validate but still confuse the runner or client parser.
• When extracting global variables, use data.variableType: "globals"; if a scope is needed, prefer data.shareScope: "PROJECT". TEAM is team-wide and may depend on paid capabilities; do not default to it unless explicitly requested.
• Always run cli-schema validate before writing and test-case get after writing.

Example:

Running And Verification

• Run the case after writing when execution behavior matters.
• Use explicit --environment for reproducibility.
• Generate a local JSON report when diagnosing runner behavior.
• Do not treat client display success as execution success; validate both saved structure and run result when required.

Common Recovery