| name | shaft-mastery |
| description | Ten distilled expert domains for working on SHAFT_ENGINE (Java test-automation framework). Load the ONE matching chapter from references/ before deep work on: Selenium/WebDriver BiDi or the recorder (selenium-bidi), Allure report internals (allure-internals), Appium/mobile toolchains (appium-mobile), Maven Central release engineering (maven-release), TestNG/JUnit listeners and forked-JVM properties (testng-lifecycle), IntelliJ Platform plugin dev (intellij-plugin), MCP protocol engineering (mcp-protocol), GitHub Actions CI forensics (ci-forensics), flaky-test stabilization and wait strategies (wait-strategies), or locator design and self-healing (locator-healing). Read only the chapter you need; each is self-contained and grounded in real SHAFT incidents.
|
SHAFT Mastery
Ten expert domains, one chapter each, in references/. Rule: identify the
domain the task actually touches, read that chapter fully, skip the rest.
Chapters encode both general expertise and SHAFT-specific incident history —
the "learned the hard way" notes are real, do not re-derive them.
| Chapter | Load when the task touches |
|---|
selenium-bidi.md | recorder, preload scripts, network capture, browser lifecycle |
allure-internals.md | report generation/patching, results JSON, verdict analysis |
appium-mobile.md | mobile recording/replay, emulators, Appium/WinAppDriver CI |
maven-release.md | versioning, Central publishing, BOM, dependency convergence |
testng-lifecycle.md | listeners, forked JVMs, properties precedence, scoped runs |
intellij-plugin.md | shaft-intellij Swing UI, tool windows, Gradle/JDK, tests |
mcp-protocol.md | shaft-mcp tools, stdio transport, workspace roots, clients |
ci-forensics.md | red CI runs, scheduled suites, workflow YAML, sharding |
wait-strategies.md | flaky tests, races, synchronization, deterministic repro |
locator-healing.md | locator choice, aria/semantic selectors, healer/doctor |
