<< All versions
Skill v1.0.1
currentAutomated scan100/100ingvarconsulting/unica/test-authoring
+3 new
──Details
PublishedAugust 11, 2026 at 07:19 AM
Content Hashsha256:fc800d11a1c5e0e7...
Git SHA9d2cfeb3a576
Bump Typepatch
──Files
Files (1 file, 3.5 KB)
SKILL.md3.5 KBactive
SKILL.md · 78 lines · 3.5 KB
version: "1.0.1" name: test-authoring description: "Проектирование и запуск тестов 1С: test-authoring, yaxunit-test и Vanessa Automation. Используй когда нужно написать тест, подобрать сценарии, запустить all/module тесты или проверить изменение через Unica runtime."
Test Authoring
MCP routing
- Preferred path: use MCP
unicatoolsunica.code.search,unica.project.map,unica.runtime.execute, and the relevantunica.*.infotools. - Use
unica.standards.searchorunica.standards.explainonly when test design depends on adevelopment-standard. Expected platform API or mechanics require aplatform-helpsource; if public MCPunicadoes not expose one, report the contract gap. - Do not call internal runtime, analyzer, or package adapters directly. They are hidden behind MCP
unica.
Workflow
- Define the behavior under test before choosing the framework: pure BSL unit, object lifecycle, form behavior, integration contract, or regression around a diagnostic.
- Search existing tests and fixtures with
unica.code.search; follow local naming, setup, teardown, and assertion style. - Prefer YaXUnit for module/unit-level BSL behavior and Vanessa Automation for UI/business scenarios that require a client.
- Build the smallest stable fixture. Avoid dependence on production data unless the user explicitly requests an integration test.
- Run
unica.runtime.executewithoperation=syntaxafter adding test code, thenoperation=testwithtestRunner=yaxunitortestRunner=va. - Report exact failing test, expected/actual behavior, and whether the failure is test setup or product behavior.
Verification gate
- For implementation plans, every stated behavior gets either an executable test,
a syntax/diagnostic check, or an explicit residual risk.
- For public API, integration, release, or metadata behavior, include impact
analysis evidence from the relevant unica.* tools before treating the test plan as complete.
- Do not call donor-specific check commands; route verification through
unica.runtime.execute, unica.code.diagnostics, and focused unica.*.info tools.
Scenario design
- Read
../../references/platform/integration-contracts.mdwhen tests verify HTTP/API/OData/JSON/XML/file-exchange behavior. - Read
../../references/platform/runtime-diagnostics.mdwhen a test is meant to reproduce a user-facing runtime failure. - Treat tests as executable debugging: one test should prove the intended user/API scenario, the failure mode, and the regression boundary.
- For API scenarios, cover success, validation error, auth error, duplicate/idempotent retry, remote timeout, and stable error semantics.
- For UI or web-client scenarios, use
operation=testfor the 1C test suite and hand a concrete autonomous URL to an external browser-testing tool when UI automation is required.
MCP examples
json
{"jsonrpc": "2.0","method": "tools/call","params": {"name": "unica.runtime.execute","arguments": {"cwd": "<workspace>","operation": "test","testRunner": "yaxunit","testScope": "module","module": "ТестДокументаЗаказКлиента","dryRun": false}}}
json
{"jsonrpc": "2.0","method": "tools/call","params": {"name": "unica.runtime.execute","arguments": {"cwd": "<workspace>","operation": "test","testRunner": "va","dryRun": false}}}