ワンクリックで
testing-and-ci
Activate this skill whenever writing tests, running linters, or committing code.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Activate this skill whenever writing tests, running linters, or committing code.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Activate this skill whenever preparing a release, bumping versions, or creating tags.
Activate this skill whenever opening Pull Requests, creating branches, or responding to GitHub issues.
Activate this skill whenever designing new features, modifying Docker compose logic, or interacting with Liferay environments.
Run FIRST before any task when workspace root lacks AGENTS.md. Creates AGENTS.md tailored to the Python environment (pyproject.toml, .venv, pytest).
Guides agents on documenting, naming, and categorizing upstream bugs/limitations in the repository's JIRA issue tracker.
Guides and scripts for developing, testing, linting, and releasing Liferay Docker Manager (LDM).
| name | testing-and-ci |
| description | Activate this skill whenever writing tests, running linters, or committing code. |
pre-commit hooks are installed locally (i.e. .git/hooks/pre-commit exists). If they are missing, you MUST run .pytest_venv/bin/pre-commit install (or ldm dev-setup) before attempting to commit code. This ensures that the local git hooks will intercept git commit and run linters like ruff format automatically, preventing unformatted code from slipping through and failing the CI Quality Gate.git commit --no-verify or bypassing quality gates. You MUST ONLY use the ./scripts/agent_push.sh "<commit message>" wrapper script, which mechanically forces the execution of pre-commit run --all-files and pytest.semgrep or detect-secrets which only run in CI), you are FORBIDDEN from using --no-verify. You MUST explicitly skip ONLY the failing hooks using the SKIP environment variable (e.g., SKIP=semgrep,detect-secrets git commit -m "..."). For formatting failures (e.g. ruff format), you MUST run the formatter locally, stage the fixed files, and try again.detect-secrets hook will fail in CI if line numbers for existing tracked secrets shift due to code changes above them (e.g., adding lines to ci.yml). When making structural changes or adding lines to files tracked in .secrets.baseline, you MUST proactively run .venv/bin/pre-commit run detect-secrets --all-files (or manually patch the line numbers in .secrets.baseline) and commit the updated baseline file to prevent CI cascade failures..ps1 and .psm1) MUST strictly use pure ASCII encoding without non-ASCII multi-byte characters. Windows PowerShell 5.1 misparses non-ASCII UTF-8 bytes on ANSI code pages. Validate via scripts/check_powershell_ascii.py.lfr-tunnel, ldm) during unit/integration tests using subprocess or os.system. All system and binary execution calls MUST be correctly mocked (@patch("ldm_core.utils.run_command") or @patch("subprocess.Popen")) to prevent triggering corporate endpoint protection tools (e.g., SentinelOne), which may detect these test invocations as malicious activity and aggressively quarantine/delete the binaries and surrounding development tools (like brew, jenv, etc.)..venv).