Vanessa_for_AI
Rules, two skills and a standard step catalog for writing Vanessa Automation Turbo Gherkin scenarios with an agent
Install
python tools/build_steps_catalog.py `
--source .\vendor\vanessa-automation `
--version 1.2.043.28 `
--commit f3a01778a14d29b38204685deea0131274d438ff `
--output knowledge/vanessa-automation/steps-library.jsonRebuilds the step catalog from local Vanessa Automation sources, PowerShell example.
This is third-party code. Review the repository files before installing.
What it does
The kit teaches the agent to write and fix Russian .feature scenarios for Vanessa Automation without MCP and without touching the 1C UI. The vanessa-project-bootstrap skill studies the repo and proposes a minimal folder structure, creating it only after approval. The vanessa-feature-authoring skill builds a scenario from the user's Vanessa recording, separates setup, action, checks and cleanup, adds waits and verifies steps against a local catalog. The step catalog is built from Vanessa Automation sources and can be rebuilt for your version with a script.
Who it is for. For 1C developers and testers automating UI tests in Vanessa Automation.
Good fit when
- You need to turn a Vanessa recording into a clean scenario with checks
- You need to diagnose a run failure and fix the scenario
- You need a structure for a new scenario project
Not a fit when
- You need the agent to launch 1C and run tests itself
- You have no recording or verified analog and the agent would have to guess UI elements
Example request
Use vanessa-feature-authoring: here is a Vanessa recording, check that the goods receipt posts and stock increasesLimitations
Requires 1C:Enterprise, Vanessa Automation, TestClient and a test database. The agent does not run scenarios: the user runs them and returns results. The bundled step catalog targets version 1.2.043.28 and does not confirm window or element names of a specific configuration. Rules and skills are CC BY 4.0 and require attribution.
How to disable. Remove the copied AGENTS.md, CLAUDE.md, the .agents/skills and .claude/skills folders with these skills, and knowledge/vanessa-automation from the project.
Security check
- The kit consists of rules, instructions and a step reference
- The agent edits only scenario files and does not connect to 1C
README in short
The README explains that the project works only with files and needs no MCP, and lists its parts: shared rules, two skills, a step catalog and its build script. It covers a quick start via GitHub Template and adding it to an existing project by copying files. A table shows how to invoke the skills in Codex, Claude Code and OpenCode. It describes how step sources are prioritized and a candid loop where a person runs the scenarios. CC BY 4.0 license, with the step catalog under BSD-3-Clause.
SKILL.md
--- name: vanessa-feature-authoring description: Создаёт, изменяет, проверяет и диагностирует русскоязычные Turbo Gherkin `.feature` для Vanessa Automation без MCP. Используй для UI-сценариев 1С, экспортных сценариев, переменных, тестовых данных, очистки, переходов окон, ожиданий, поиска шагов в локальном каталоге и разбора runtime-ошибок. Требует запись Vanessa Automation или runtime-проверенный аналог и бизнес-цель пользователя. --- # Авторинг сценариев Vanessa Automation Не подключай MCP. Работай с файлами проекта, локальным каталогом шагов, записью пользователя и runtime-диагностикой. ## Проверь вход До изменения UI-шагов получи: 1. бизнес-цель; 2. ожидаемые проверки результата; 3. запись Vanessa Automation либо runtime-проверенный аналог. Если структуры проекта нет, сначала используй `vanessa-project-bootstrap` и дождись согласования. Если не подтверждена конкретная UI-деталь, запроси минимальный фрагмент записи или диагностики до редактирования затронутых шагов. ## Собери контекст 1. Прочитай `AGENTS.md`, корневой README и ближайший README сценария. 2. Найди похожие исполняемые и экспортные сценарии. 3. Проверь точные имена вызываемых `@ExportScenarios`. 4. Определи версию Vanessa Automation и подходящий каталог шагов. 5. Составь порядок зависимостей и очистки. Читайте references только по потребности: - подтверждение UI-фактов — [references/ui-evidence.md](references/ui-evidence.md); - состояния окон и ожидания — [references/window-state-and-waits.md](references/window-state-and-waits.md); - переменные и экспортные сценарии — [references/variables-and-exports.md](references/variables-and-exports.md); - тестовые данные и очистка — [references/data-and-cleanup.md](references/data-and-cleanup.md); - ожидаемые ошибки и сообщения пользователю — [references/negative-messages.md](references/negative-messages.md); - условные блоки `Если ... Тогда` — [references/conditional-blocks.md](references/conditional-blocks.md); - каталог шагов — [references/step-catalog.md](references/step-catalog.md); - диагностика и финальная проверка — [references/diagnostics-and-review.md](references/diagnostics-and-review.md). ## Спроектируй сценарий 1. Отдели подготовку, действие, независимые бизнес-проверки и очистку. 2. Для каждого действия, меняющего окно, запиши трассу: `действие -> ожидаемое активное окно -> источник факта`
FAQ
Why is a Vanessa recording needed?
The skill does not invent window titles or element names. Without a recording or verified scenario it asks for a minimal diagnostic snippet.
How do I get a catalog for my version?
Run tools/build_steps_catalog.py with a path to a local Vanessa Automation checkout and your version.
Related
The official MCP server debugger: web UI, automation CLI and terminal UI in one package
SonarSource's official MCP server: quality and security issues, quality gates and code analysis from SonarQube Server and Cloud
Playwright CLI
playwright-cli
Microsoft's official Playwright CLI with an agent skill: drive a browser through short commands without heavy MCP schemas
Open-source AI agent for pull request review: descriptions, feedback and improvement suggestions on GitHub, GitLab and more