[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AJS5TDSYHAQHVT8VWCT5C0":3,"skill-same:01M7AJS5TDSYHAQHVT8VWCT5C0":49},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":25,"commitSha":26,"skillDirectory":27,"forksCount":28,"repositoryHtmlUrl":29,"repositoryHomepageUrl":30,"ownerHtmlUrl":31,"ownerBlogUrl":32,"ownerDescription":33,"updatedAt":34,"install":35,"resources":48},"01M7AJS5TDSYHAQHVT8VWCT5C0","vibey-quality-gates",null,"The 7-gate CI sweep (ruff check, ruff format, mypy, per-layer coverage × 4, lint-imports, bandit, pip-audit), what each gate catches, the other CI jobs, and the workspace tenants' own checks.","the-vibey-project\u002Fvibey","the-vibey-project","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F329298758?v=4","01M7AJHGZ9722CTJ4JNQZVRH0Q","the vibey project","01M7AJHGZVMG2K7TGXQ6XFHVG3",5,"MIT","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002F.claude\u002Fskills\u002Fvibey-quality-gates\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: vibey-quality-gates\ndescription: The 7-gate CI sweep (ruff check, ruff format, mypy, per-layer coverage × 4, lint-imports, bandit, pip-audit), what each gate catches, the other CI jobs, and the workspace tenants' own checks.\nallowed-tools: Bash\n---\n\n# vibey quality gates\n\nCI runs a 7-gate sweep over the root `vibey` package, split over two jobs in\n`.github\u002Fworkflows\u002Fci.yml`: `lint` (named \"Lint and static gates\": Gates 1, 2, 3, 3b and 5\nwith the corpus index and TypeScript artifact checks) and `gates` (the test run, Gate 4,\nbandit and pip-audit). **All seven must pass** — together with the other CI\njobs below — before a PR can merge. Run them locally with\n`pre-commit run --all-files --hook-stage pre-push` (a plain\n`pre-commit run --all-files` runs only the ruff hooks) or individually:\n\n## Gate 1: ruff check\n\nLinter. Catches unused imports, undefined names, syntax errors, and common\nanti-patterns.\n\n```bash\nuv run ruff check .\n```\n\nRoot ruff covers `src\u002F**\u002F*.py` and `tests\u002F**\u002F*.py`, so it also lints the\nworkspace tenants under `src\u002Fvibey_runners\u002F` and `src\u002Fvibey_tools\u002F`.\n\n## Gate 2: ruff format --check\n\nFormatter. Ensures consistent code style across the repo. Fails if any file\nwould be reformatted.\n\n```bash\nuv run ruff format --check .\n```\n\nTo fix: `uv run ruff format .`\n\n## Gate 3: mypy --strict\n\nType checker. Enforces `--strict` mode across `src\u002Fvibey\u002F`. Catches type\nerrors, missing type annotations, and `Any` usage.\n\n```bash\nuv run mypy --strict src\u002Fvibey\n```\n\nGate 3b runs the same checker over the other declared trees (`scripts\u002F`, the storm\ntools and the tenants), read from `[[tool.vibey.static_gates]]`:\n\n```bash\nuv run python scripts\u002Fstatic_gates.py\n```\n\n## Gate 4: pytest per-layer coverage (100%)\n\nOne test run produces a combined `.coverage` file; four per-layer reports\nenforce 100% branch coverage. Each must pass or the build fails.\n\n```bash\nuv run pytest -q -p no:cacheprovider --cov=vibey --cov-branch --cov-report=\nuv run coverage report --include='src\u002Fvibey\u002Fdomain\u002F*' --fail-under=100\nuv run coverage report --include='src\u002Fvibey\u002Fapplication\u002F*' --fail-under=100\nuv run coverage report --include='src\u002Fvibey\u002Finfrastructure\u002F*' --fail-under=100\nuv run coverage report --include='src\u002Fvibey\u002Fcli\u002F*' --fail-under=100\n```\n\n**Why per-layer, not aggregate?** Because `domain\u002F` is the safety-critical\nlayer and must never drop below 100%. An aggregate gate would allow a 95%\ndomain to hide behind a 100% infrastructure. See ADR-0023.\n\nThe test run needs a reachable Postgres even for a subset: `tests\u002Fconftest.py`\nclones a migrated template database per xdist worker at session start. See the\n`vibey-testing` skill.\n\n## Gate 5: lint-imports\n\nOnion architecture enforcer. Verifies dependencies point inward only, and\nthat `domain\u002F` imports nothing but stdlib.\n\n```bash\nuv run lint-imports\n```\n\nIf it fails, the error names the contract and the violating import. See\n`.importlinter` for the contracts: `onion-layers`, `domain-independence`,\n`application-independence`, and the declare-only contracts that keep each\n`interfaces\u002F` package from importing its consumers (`interfaces-declare-only`\nfor `application\u002Finterfaces`, `infrastructure-interfaces-declare-only`,\n`vibey-gh-interfaces-declare-only`). The declare-only rule applies to every\n`interfaces\u002F` package. See ADR-0016.\n\nSee the `vibey-architecture` skill before fixing a violation.\n\n## Gate 6: bandit\n\nSecurity linter. Scans for common security issues: hardcoded passwords, SQL\ninjection, insecure temp files, weak crypto, shell=True, etc.\n\n```bash\nuv run bandit -q -r src\u002Fvibey\n```\n\n## Gate 7: pip-audit\n\nDependency vulnerability scanner. Checks for known CVEs in pinned\ndependencies.\n\n```bash\nuv run pip-audit\n```\n\nThe npm lockfiles (the root workspace and `clients\u002Fapp`) get the same gate from\n`vibey-gh advisory-check` in the `@vibey\u002Fcore`, `app` and the two `VS Code extension` jobs:\n`npm audit` at `[advisories] audit_level` (high), less the declared exceptions in\n`.github\u002Fadvisory-exceptions.toml`. An exception is never a silent skip: it names one\nadvisory, package and workspace, records why the vulnerable code is not reached, and\nexpires within `max_exception_days` (30). Every run prints each exception it honours and\nuntil when, and fails when one has expired, matches nothing (remove it), or its package has\na patched release (take the fix). The file is a protected path, so accepting a risk takes the\ncode owner's approval and a human merge. An outage of the advisory service itself is handled\nseparately: `on_unavailable = \"fail\"` is the default, and this repository sets `\"warn\"`, so a\nnonresponsive endpoint prints `::warning::` naming the workspace and the reason and continues\nunaudited; a real advisory, a malformed report or a missing lockfile still fails.\n`SECURITY.md`, \"Dependency advisories\", has the rest.\n\n## Hook diet\n\nThe seven gates are split across git hook stages so commits stay fast\n(`.pre-commit-config.yaml`):\n\n**Pre-commit stage** (runs on every `git commit`):\n- ruff (`--fix`) + ruff format, on changed files\n\n**Pre-push stage** (runs on every `git push`):\n- The plain parallel test suite (`uv run pytest -q -p no:cacheprovider`) — kept\n  so a test failure reports as itself rather than as a coverage-gate failure\n- mypy --strict\n- lint-imports\n- Per-layer 100% coverage gates (pytest --cov + four reports)\n- bandit\n- pip-audit\n\n**Commit-msg stage**: Conventional Commits enforcement.\n\n**Pushing**: push through the push gate, never with a bare `git push`. Run\n`python3 \u003Cstorm>\u002Ftools\u002Fpush_gate.py run -- git push origin HEAD:\u003Cbranch>`, or from a\ncheckout with no storm, set `VIBEY_PUSH_LOCK=\u003Cdir>` and run the tracked\n`docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fpush_gate.py`. It lets one pre-push run happen at a time\nacross every lane on the machine. Exit 124 (`reaped: hang`) means the reaper stopped a hung\ngate run; it is not a test failure. CONTRIBUTING.md, \"Pushing in this repository\", has the\nrest.\n\nInstall all three hook types once, then the provenance hooks, with one idempotent\ncommand (the order matters, because `pre-commit install` refuses to run while\n`core.hooksPath` is set, and the command does it in that order):\n\n```bash\nuv run --extra dev vibey-gh setup\n```\n\n`vibey-gh install` points `core.hooksPath` at `.githooks`. The tracked shims there\n(`pre-commit`, `commit-msg.local`, `pre-push.local`) chain back to the framework\nthrough `git rev-parse --git-common-dir`, so one install covers the main checkout\nand every linked worktree (`git worktree add`). If a shim prints\n`warning: the pre-commit framework's \u003Cstage> hook is not installed`, the commit or\npush went ahead but none of that stage's gates ran. Run them by hand before you push.\n\n## The other CI jobs\n\n`ci.yml` runs fourteen jobs on every push and PR to `develop`\u002F`main`. The seven\nbelow are the Python and deployment gates; the others are `noloss` (the no-loss\nproperty suite at 10,000 examples), `postgres-compatibility` (the database suite on\nPostgreSQL 14 to 18), `krypton-app`, `vibey-core` (`@vibey\u002Fcore`), `app` (the krypton\napp, mobile and web), the two `VS Code extension` jobs (Ubuntu and macOS) and `desktop` (the krypton\ndesktop client on Ubuntu, Arch, Fedora and macOS, where it also bundles, verifies, launches and\nimages the app):\n\n| Job | What it checks |\n|---|---|\n| `uv-lock` | `uv lock --check`. `gates`, `tools` and `tools-lint` depend on it. The lock carries vibey's own version, so a version bump without `uv lock` fails here first. |\n| `gates` | The seven gates above, against a `postgres:17` service. `postgres-compatibility` additionally runs the database suite on PostgreSQL 14, 15, 16, 17, and 18. |\n| `tools` | Each absorbed tenant's own suite on its own Python floors, plus, on its floor row, its own static gates from the row's `static` key: its mypy, `lint-imports` and bandit. codexloop and vibey-skills also run their own strict docs builds (the `docs` key) (ADR-0022). |\n| `tools-lint` | vibey-gh's own linters and its managed-automation drift check. |\n| `image` | Builds `deploy\u002Fdocker\u002FDockerfile` for amd64 and arm64 and asserts each `Image contract - …` step: the entrypoint runs, it runs as non-root uid 10001, it has no compiler\u002Fuv\u002Fpip, migrations ship in the image, and every console script is on PATH. |\n| `chart` | Render-only: `deploy\u002Fhelm\u002Fgolden\u002Frender.sh` runs `helm lint --strict` and `helm template` for each profile (defaults, `ollama.enabled`, the GPU + gptossloop wiring with qwenloop switched on beside it, and the KEDA query unbound and bound to a project) and diffs each render against its committed golden under `deploy\u002Fhelm\u002Fgolden\u002F`, with helm pinned. After an intended chart change, regenerate with `deploy\u002Fhelm\u002Fgolden\u002Frender.sh --update`. |\n| `cluster-smoke` | Helm install of `deploy\u002Fhelm\u002Fvibey` on minikube and four cluster contracts: a projectless worker parks instead of crash-looping, the worker picks up a project created in-cluster, the KEDA ScaledObject reconciles against real Postgres, and a worker drains promptly on SIGTERM (ADR-0025, ADR-0026). |\n\nOther workflows also gate a merge: `provenance.yml` (runs on every push and PR),\nand `pr-evaluate.yml`, which re-evaluates a PR each time `CI` or `Provenance`\ncompletes and dispatches `pr-review.yml` for the review (`[pr_automation] scan_workflows` in `.vibey-gh.toml`) and hands ready\nPRs to `merge-train.yml` (see the `vibey-releasing` skill).\n\n## Workspace tenants: their own checks\n\nThe tenants under `src\u002Fvibey_runners\u002F` and `src\u002Fvibey_tools\u002F` keep the gates they\narrived with (ADR-0021, ADR-0022). The root `mypy`, `bandit` and coverage gates\ncover `src\u002Fvibey` only. Run a tenant's checks from its own directory, with plain\npip and a Python at its floor — every tenant now shares the 3.12 floor, so the\nmatrix exercises the common 3.12–3.14 range. Plain pip knows nothing\nabout `[tool.uv.sources] workspace = true`, so a tenant that needs a sibling\ninstalls that sibling **from the tree first** -- no family package may be\nrequested from an index, because none of them is published any more (ADR-0037).\nThese are the `tools` and `tools-lint` commands, verbatim:\n\n```bash\n# vibey-gh (CI: Python 3.12, 3.13, 3.14; 100% branch floor in its own addopts)\ncd src\u002Fvibey_tools\u002Fgh\npip install -e \".[dev]\"\npython -m pytest -q\npython -m black --check vibey_gh test\nisort --check-only vibey_gh test\npython -m mypy vibey_gh\n\n# vibey-skills (CI: Python 3.12, 3.13, 3.14)\ncd src\u002Fvibey_tools\u002Fskills\npip install -e .\npython3 tools\u002Fvalidate_manifests.py && python3 tools\u002Fcheck_links.py \\\n  && PYTHONPATH=src python3 -m unittest discover -s tests\n\n# vibey-bootstrap (CI: Python 3.12, 3.13, 3.14). It imports vibey_gh, so install that first.\ncd src\u002Fvibey_tools\u002Fbootstrap\npip install -e ..\u002Fgh && pip install -e \".[test,all]\"\npytest test\u002F -m \"not integration\" --cov=vibey_bootstrap --cov-report=term\n\n# claudeloop (CI: Python 3.12, 3.13, 3.14 -- the common library floor)\ncd src\u002Fvibey_runners\u002Fclaude\npip install -e ..\u002Fcommon && pip install -e \".[dev]\"\npytest tests\u002Fdomain --cov=claudeloop.domain --cov-branch --cov-report=term-missing --cov-fail-under=100\npytest tests\u002Fapplication --cov=claudeloop.application --cov-branch --cov-report=term-missing --cov-fail-under=100\npytest tests\u002Finfrastructure -n auto --maxprocesses=8 --cov=claudeloop.infrastructure --cov-branch --cov-report=term-missing --cov-fail-under=100\npytest tests\u002Fcli -n auto --maxprocesses=8 --cov=claudeloop.cli --cov-branch --cov-report=term-missing --cov-fail-under=100\n\n# codexloop (CI: Python 3.12, 3.13).\ncd src\u002Fvibey_runners\u002Fcodex\npip install -e ..\u002Fcommon && pip install -e \".[dev]\"\npytest -q --cov=codexloop --cov-branch --cov-report=\ncoverage report --include='src\u002Fcodexloop\u002Fdomain\u002F*' --fail-under=100\ncoverage report --include='src\u002Fcodexloop\u002Fapplication\u002F*' --fail-under=100\ncoverage report --include='src\u002Fcodexloop\u002Finfrastructure\u002F*' --fail-under=100\ncoverage report --include='src\u002Fcodexloop\u002Fcli\u002F*' --fail-under=100\n\n# qwenloop (CI: Python 3.12, 3.13, 3.14; its floor is already in its addopts)\ncd src\u002Fvibey_runners\u002Fqwen\npip install -e \".[dev]\"\npython -m pytest -q\n```\n\nvibey-gh's suite drives real git history and shells out to `gh`, so it needs\nboth on `PATH` and a full clone -- a shallow one fails. It does **not** need\n`uv`: the single test that drives a real `uv lock` is marked `network`, below.\n\nTests that leave the machine (a real package index, a real remote) carry\n`@pytest.mark.network` and are **skipped unless `VIBEY_GH_NETWORK_TESTS=1`**.\nMarker and variable are two halves of one mechanism -- the marker declares, and\n`test\u002Fconftest.py` does the skipping -- so a plain offline `python -m pytest`\nreaches the package's 100% branch floor with no flag to remember, and a run that\nreports `4 skipped` is correct rather than degraded. Every branch a `network`\ntest touches must also be reachable offline, or that floor fails. Where an index\nis reachable, ask for them by name:\n\n```bash\ncd src\u002Fvibey_tools\u002Fgh\nVIBEY_GH_NETWORK_TESTS=1 python -m pytest -q -m network --no-cov\n```\n\nCI runs that on one matrix row with `continue-on-error`: it reports on somebody\nelse's service, so it must never gate a merge.\n\nThe runners (`src\u002Fvibey_runners\u002F{claude,codex,qwen}`) carry their own\n`tests\u002F`, ruff, mypy and import-linter configuration in their `pyproject.toml`.\nSince 2026-09-15 the root `tools` matrix runs all three, one row per runner and Python\nversion (read the rows in `ci.yml`), so a runner regression turns the `tools` job red. It reports rather than blocks: `tools`\nis not named in `required_checks` in `.vibey-gh.toml`, and no tools row ever has been.\nRead a runner's `[tool.pytest.ini_options]` before running anything by hand. Only qwenloop's\naddopts carry a coverage floor. The other two deliberately leave `--cov=` and\n`--cov-fail-under` out, because pytest-cov unions every `--cov=` it sees and one baked\nin there would widen the explicitly-scoped per-layer runs back out to the whole package\n-- so for them a bare `pytest` runs the tests but enforces no floor. Use the commands\nabove, which are their own.\n\n## Where work lives, and how often it is saved\n\nKeep every clone and worktree on storage a reboot keeps (sub-doctrine 10.h, ADR-0057). Never\nuse `\u002Ftmp`, `\u002Fprivate\u002Ftmp`, `\u002Fvar\u002Ftmp`, `\u002Fvar\u002Ffolders`, `\u002Fdev\u002Fshm`, `\u002Frun\u002Fuser` or `$TMPDIR`:\nthe OS empties them. On 2026-09-24 a reboot emptied `\u002Fprivate\u002Ftmp` mid-storm and took every\nuncommitted worktree, measurement and draft with it. Worktrees go in the storm home:\n`VIBEY_STORM_HOME`, else `~\u002Fgit\u002Fvibey-storm` on macOS, `$XDG_DATA_HOME\u002Fvibey\u002Fstorm` (else\n`~\u002F.local\u002Fshare\u002Fvibey\u002Fstorm`) on Linux.\n\n```bash\npython3 docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fstorm_durability.py status   # durable or not\ngit worktree add \"$(python3 docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fstorm_durability.py worktree fix-x)\" \\\n  -b fix\u002Fx origin\u002Fdevelop\n```\n\n- Commit as soon as a change is coherent, not when it is finished.\n- Push work in progress to a draft PR (`gh pr create --draft`) at least every 30–45 minutes;\n  the merge train never merges a draft, and the pre-push gates still run.\n- A long measurement writes each step as it finishes and resumes (`StepJournal` in\n  `docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fstorm_checkpoint.py`).\n- The storm tools refuse volatile storage with exit 78 and name the key to change.\n  CONTRIBUTING.md, \"Where your work lives\", has the rest.\n\n## What each gate catches\n\n| Gate | Catches |\n|---|---|\n| ruff check | Unused imports, undefined names, syntax errors, anti-patterns |\n| ruff format | Inconsistent code style |\n| mypy --strict | Type errors, missing annotations, `Any` usage |\n| pytest (per-layer 100%) | Untested branches, logic errors, regressions |\n| lint-imports | Onion violations, forbidden imports |\n| bandit | Security issues, hardcoded secrets, insecure patterns |\n| pip-audit | Dependency CVEs |\n","# vibey quality gates\n\nCI runs a 7-gate sweep over the root `vibey` package, split over two jobs in\n`.github\u002Fworkflows\u002Fci.yml`: `lint` (named \"Lint and static gates\": Gates 1, 2, 3, 3b and 5\nwith the corpus index and TypeScript artifact checks) and `gates` (the test run, Gate 4,\nbandit and pip-audit). **All seven must pass** — together with the other CI\njobs below — before a PR can merge. Run them locally with\n`pre-commit run --all-files --hook-stage pre-push` (a plain\n`pre-commit run --all-files` runs only the ruff hooks) or individually:\n\n## Gate 1: ruff check\n\nLinter. Catches unused imports, undefined names, syntax errors, and common\nanti-patterns.\n\n```bash\nuv run ruff check .\n```\n\nRoot ruff covers `src\u002F**\u002F*.py` and `tests\u002F**\u002F*.py`, so it also lints the\nworkspace tenants under `src\u002Fvibey_runners\u002F` and `src\u002Fvibey_tools\u002F`.\n\n## Gate 2: ruff format --check\n\nFormatter. Ensures consistent code style across the repo. Fails if any file\nwould be reformatted.\n\n```bash\nuv run ruff format --check .\n```\n\nTo fix: `uv run ruff format .`\n\n## Gate 3: mypy --strict\n\nType checker. Enforces `--strict` mode across `src\u002Fvibey\u002F`. Catches type\nerrors, missing type annotations, and `Any` usage.\n\n```bash\nuv run mypy --strict src\u002Fvibey\n```\n\nGate 3b runs the same checker over the other declared trees (`scripts\u002F`, the storm\ntools and the tenants), read from `[[tool.vibey.static_gates]]`:\n\n```bash\nuv run python scripts\u002Fstatic_gates.py\n```\n\n## Gate 4: pytest per-layer coverage (100%)\n\nOne test run produces a combined `.coverage` file; four per-layer reports\nenforce 100% branch coverage. Each must pass or the build fails.\n\n```bash\nuv run pytest -q -p no:cacheprovider --cov=vibey --cov-branch --cov-report=\nuv run coverage report --include='src\u002Fvibey\u002Fdomain\u002F*' --fail-under=100\nuv run coverage report --include='src\u002Fvibey\u002Fapplication\u002F*' --fail-under=100\nuv run coverage report --include='src\u002Fvibey\u002Finfrastructure\u002F*' --fail-under=100\nuv run coverage report --include='src\u002Fvibey\u002Fcli\u002F*' --fail-under=100\n```\n\n**Why per-layer, not aggregate?** Because `domain\u002F` is the safety-critical\nlayer and must never drop below 100%. An aggregate gate would allow a 95%\ndomain to hide behind a 100% infrastructure. See ADR-0023.\n\nThe test run needs a reachable Postgres even for a subset: `tests\u002Fconftest.py`\nclones a migrated template database per xdist worker at session start. See the\n`vibey-testing` skill.\n\n## Gate 5: lint-imports\n\nOnion architecture enforcer. Verifies dependencies point inward only, and\nthat `domain\u002F` imports nothing but stdlib.\n\n```bash\nuv run lint-imports\n```\n\nIf it fails, the error names the contract and the violating import. See\n`.importlinter` for the contracts: `onion-layers`, `domain-independence`,\n`application-independence`, and the declare-only contracts that keep each\n`interfaces\u002F` package from importing its consumers (`interfaces-declare-only`\nfor `application\u002Finterfaces`, `infrastructure-interfaces-declare-only`,\n`vibey-gh-interfaces-declare-only`). The declare-only rule applies to every\n`interfaces\u002F` package. See ADR-0016.\n\nSee the `vibey-architecture` skill before fixing a violation.\n\n## Gate 6: bandit\n\nSecurity linter. Scans for common security issues: hardcoded passwords, SQL\ninjection, insecure temp files, weak crypto, shell=True, etc.\n\n```bash\nuv run bandit -q -r src\u002Fvibey\n```\n\n## Gate 7: pip-audit\n\nDependency vulnerability scanner. Checks for known CVEs in pinned\ndependencies.\n\n```bash\nuv run pip-audit\n```\n\nThe npm lockfiles (the root workspace and `clients\u002Fapp`) get the same gate from\n`vibey-gh advisory-check` in the `@vibey\u002Fcore`, `app` and the two `VS Code extension` jobs:\n`npm audit` at `[advisories] audit_level` (high), less the declared exceptions in\n`.github\u002Fadvisory-exceptions.toml`. An exception is never a silent skip: it names one\nadvisory, package and workspace, records why the vulnerable code is not reached, and\nexpires within `max_exception_days` (30). Every run prints each exception it honours and\nuntil when, and fails when one has expired, matches nothing (remove it), or its package has\na patched release (take the fix). The file is a protected path, so accepting a risk takes the\ncode owner's approval and a human merge. An outage of the advisory service itself is handled\nseparately: `on_unavailable = \"fail\"` is the default, and this repository sets `\"warn\"`, so a\nnonresponsive endpoint prints `::warning::` naming the workspace and the reason and continues\nunaudited; a real advisory, a malformed report or a missing lockfile still fails.\n`SECURITY.md`, \"Dependency advisories\", has the rest.\n\n## Hook diet\n\nThe seven gates are split across git hook stages so commits stay fast\n(`.pre-commit-config.yaml`):\n\n**Pre-commit stage** (runs on every `git commit`):\n- ruff (`--fix`) + ruff format, on changed files\n\n**Pre-push stage** (runs on every `git push`):\n- The plain parallel test suite (`uv run pytest -q -p no:cacheprovider`) — kept\n  so a test failure reports as itself rather than as a coverage-gate failure\n- mypy --strict\n- lint-imports\n- Per-layer 100% coverage gates (pytest --cov + four reports)\n- bandit\n- pip-audit\n\n**Commit-msg stage**: Conventional Commits enforcement.\n\n**Pushing**: push through the push gate, never with a bare `git push`. Run\n`python3 \u003Cstorm>\u002Ftools\u002Fpush_gate.py run -- git push origin HEAD:\u003Cbranch>`, or from a\ncheckout with no storm, set `VIBEY_PUSH_LOCK=\u003Cdir>` and run the tracked\n`docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fpush_gate.py`. It lets one pre-push run happen at a time\nacross every lane on the machine. Exit 124 (`reaped: hang`) means the reaper stopped a hung\ngate run; it is not a test failure. CONTRIBUTING.md, \"Pushing in this repository\", has the\nrest.\n\nInstall all three hook types once, then the provenance hooks, with one idempotent\ncommand (the order matters, because `pre-commit install` refuses to run while\n`core.hooksPath` is set, and the command does it in that order):\n\n```bash\nuv run --extra dev vibey-gh setup\n```\n\n`vibey-gh install` points `core.hooksPath` at `.githooks`. The tracked shims there\n(`pre-commit`, `commit-msg.local`, `pre-push.local`) chain back to the framework\nthrough `git rev-parse --git-common-dir`, so one install covers the main checkout\nand every linked worktree (`git worktree add`). If a shim prints\n`warning: the pre-commit framework's \u003Cstage> hook is not installed`, the commit or\npush went ahead but none of that stage's gates ran. Run them by hand before you push.\n\n## The other CI jobs\n\n`ci.yml` runs fourteen jobs on every push and PR to `develop`\u002F`main`. The seven\nbelow are the Python and deployment gates; the others are `noloss` (the no-loss\nproperty suite at 10,000 examples), `postgres-compatibility` (the database suite on\nPostgreSQL 14 to 18), `krypton-app`, `vibey-core` (`@vibey\u002Fcore`), `app` (the krypton\napp, mobile and web), the two `VS Code extension` jobs (Ubuntu and macOS) and `desktop` (the krypton\ndesktop client on Ubuntu, Arch, Fedora and macOS, where it also bundles, verifies, launches and\nimages the app):\n\n| Job | What it checks |\n|---|---|\n| `uv-lock` | `uv lock --check`. `gates`, `tools` and `tools-lint` depend on it. The lock carries vibey's own version, so a version bump without `uv lock` fails here first. |\n| `gates` | The seven gates above, against a `postgres:17` service. `postgres-compatibility` additionally runs the database suite on PostgreSQL 14, 15, 16, 17, and 18. |\n| `tools` | Each absorbed tenant's own suite on its own Python floors, plus, on its floor row, its own static gates from the row's `static` key: its mypy, `lint-imports` and bandit. codexloop and vibey-skills also run their own strict docs builds (the `docs` key) (ADR-0022). |\n| `tools-lint` | vibey-gh's own linters and its managed-automation drift check. |\n| `image` | Builds `deploy\u002Fdocker\u002FDockerfile` for amd64 and arm64 and asserts each `Image contract - …` step: the entrypoint runs, it runs as non-root uid 10001, it has no compiler\u002Fuv\u002Fpip, migrations ship in the image, and every console script is on PATH. |\n| `chart` | Render-only: `deploy\u002Fhelm\u002Fgolden\u002Frender.sh` runs `helm lint --strict` and `helm template` for each profile (defaults, `ollama.enabled`, the GPU + gptossloop wiring with qwenloop switched on beside it, and the KEDA query unbound and bound to a project) and diffs each render against its committed golden under `deploy\u002Fhelm\u002Fgolden\u002F`, with helm pinned. After an intended chart change, regenerate with `deploy\u002Fhelm\u002Fgolden\u002Frender.sh --update`. |\n| `cluster-smoke` | Helm install of `deploy\u002Fhelm\u002Fvibey` on minikube and four cluster contracts: a projectless worker parks instead of crash-looping, the worker picks up a project created in-cluster, the KEDA ScaledObject reconciles against real Postgres, and a worker drains promptly on SIGTERM (ADR-0025, ADR-0026). |\n\nOther workflows also gate a merge: `provenance.yml` (runs on every push and PR),\nand `pr-evaluate.yml`, which re-evaluates a PR each time `CI` or `Provenance`\ncompletes and dispatches `pr-review.yml` for the review (`[pr_automation] scan_workflows` in `.vibey-gh.toml`) and hands ready\nPRs to `merge-train.yml` (see the `vibey-releasing` skill).\n\n## Workspace tenants: their own checks\n\nThe tenants under `src\u002Fvibey_runners\u002F` and `src\u002Fvibey_tools\u002F` keep the gates they\narrived with (ADR-0021, ADR-0022). The root `mypy`, `bandit` and coverage gates\ncover `src\u002Fvibey` only. Run a tenant's checks from its own directory, with plain\npip and a Python at its floor — every tenant now shares the 3.12 floor, so the\nmatrix exercises the common 3.12–3.14 range. Plain pip knows nothing\nabout `[tool.uv.sources] workspace = true`, so a tenant that needs a sibling\ninstalls that sibling **from the tree first** -- no family package may be\nrequested from an index, because none of them is published any more (ADR-0037).\nThese are the `tools` and `tools-lint` commands, verbatim:\n\n```bash\n# vibey-gh (CI: Python 3.12, 3.13, 3.14; 100% branch floor in its own addopts)\ncd src\u002Fvibey_tools\u002Fgh\npip install -e \".[dev]\"\npython -m pytest -q\npython -m black --check vibey_gh test\nisort --check-only vibey_gh test\npython -m mypy vibey_gh\n\n# vibey-skills (CI: Python 3.12, 3.13, 3.14)\ncd src\u002Fvibey_tools\u002Fskills\npip install -e .\npython3 tools\u002Fvalidate_manifests.py && python3 tools\u002Fcheck_links.py \\\n  && PYTHONPATH=src python3 -m unittest discover -s tests\n\n# vibey-bootstrap (CI: Python 3.12, 3.13, 3.14). It imports vibey_gh, so install that first.\ncd src\u002Fvibey_tools\u002Fbootstrap\npip install -e ..\u002Fgh && pip install -e \".[test,all]\"\npytest test\u002F -m \"not integration\" --cov=vibey_bootstrap --cov-report=term\n\n# claudeloop (CI: Python 3.12, 3.13, 3.14 -- the common library floor)\ncd src\u002Fvibey_runners\u002Fclaude\npip install -e ..\u002Fcommon && pip install -e \".[dev]\"\npytest tests\u002Fdomain --cov=claudeloop.domain --cov-branch --cov-report=term-missing --cov-fail-under=100\npytest tests\u002Fapplication --cov=claudeloop.application --cov-branch --cov-report=term-missing --cov-fail-under=100\npytest tests\u002Finfrastructure -n auto --maxprocesses=8 --cov=claudeloop.infrastructure --cov-branch --cov-report=term-missing --cov-fail-under=100\npytest tests\u002Fcli -n auto --maxprocesses=8 --cov=claudeloop.cli --cov-branch --cov-report=term-missing --cov-fail-under=100\n\n# codexloop (CI: Python 3.12, 3.13).\ncd src\u002Fvibey_runners\u002Fcodex\npip install -e ..\u002Fcommon && pip install -e \".[dev]\"\npytest -q --cov=codexloop --cov-branch --cov-report=\ncoverage report --include='src\u002Fcodexloop\u002Fdomain\u002F*' --fail-under=100\ncoverage report --include='src\u002Fcodexloop\u002Fapplication\u002F*' --fail-under=100\ncoverage report --include='src\u002Fcodexloop\u002Finfrastructure\u002F*' --fail-under=100\ncoverage report --include='src\u002Fcodexloop\u002Fcli\u002F*' --fail-under=100\n\n# qwenloop (CI: Python 3.12, 3.13, 3.14; its floor is already in its addopts)\ncd src\u002Fvibey_runners\u002Fqwen\npip install -e \".[dev]\"\npython -m pytest -q\n```\n\nvibey-gh's suite drives real git history and shells out to `gh`, so it needs\nboth on `PATH` and a full clone -- a shallow one fails. It does **not** need\n`uv`: the single test that drives a real `uv lock` is marked `network`, below.\n\nTests that leave the machine (a real package index, a real remote) carry\n`@pytest.mark.network` and are **skipped unless `VIBEY_GH_NETWORK_TESTS=1`**.\nMarker and variable are two halves of one mechanism -- the marker declares, and\n`test\u002Fconftest.py` does the skipping -- so a plain offline `python -m pytest`\nreaches the package's 100% branch floor with no flag to remember, and a run that\nreports `4 skipped` is correct rather than degraded. Every branch a `network`\ntest touches must also be reachable offline, or that floor fails. Where an index\nis reachable, ask for them by name:\n\n```bash\ncd src\u002Fvibey_tools\u002Fgh\nVIBEY_GH_NETWORK_TESTS=1 python -m pytest -q -m network --no-cov\n```\n\nCI runs that on one matrix row with `continue-on-error`: it reports on somebody\nelse's service, so it must never gate a merge.\n\nThe runners (`src\u002Fvibey_runners\u002F{claude,codex,qwen}`) carry their own\n`tests\u002F`, ruff, mypy and import-linter configuration in their `pyproject.toml`.\nSince 2026-09-15 the root `tools` matrix runs all three, one row per runner and Python\nversion (read the rows in `ci.yml`), so a runner regression turns the `tools` job red. It reports rather than blocks: `tools`\nis not named in `required_checks` in `.vibey-gh.toml`, and no tools row ever has been.\nRead a runner's `[tool.pytest.ini_options]` before running anything by hand. Only qwenloop's\naddopts carry a coverage floor. The other two deliberately leave `--cov=` and\n`--cov-fail-under` out, because pytest-cov unions every `--cov=` it sees and one baked\nin there would widen the explicitly-scoped per-layer runs back out to the whole package\n-- so for them a bare `pytest` runs the tests but enforces no floor. Use the commands\nabove, which are their own.\n\n## Where work lives, and how often it is saved\n\nKeep every clone and worktree on storage a reboot keeps (sub-doctrine 10.h, ADR-0057). Never\nuse `\u002Ftmp`, `\u002Fprivate\u002Ftmp`, `\u002Fvar\u002Ftmp`, `\u002Fvar\u002Ffolders`, `\u002Fdev\u002Fshm`, `\u002Frun\u002Fuser` or `$TMPDIR`:\nthe OS empties them. On 2026-09-24 a reboot emptied `\u002Fprivate\u002Ftmp` mid-storm and took every\nuncommitted worktree, measurement and draft with it. Worktrees go in the storm home:\n`VIBEY_STORM_HOME`, else `~\u002Fgit\u002Fvibey-storm` on macOS, `$XDG_DATA_HOME\u002Fvibey\u002Fstorm` (else\n`~\u002F.local\u002Fshare\u002Fvibey\u002Fstorm`) on Linux.\n\n```bash\npython3 docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fstorm_durability.py status   # durable or not\ngit worktree add \"$(python3 docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fstorm_durability.py worktree fix-x)\" \\\n  -b fix\u002Fx origin\u002Fdevelop\n```\n\n- Commit as soon as a change is coherent, not when it is finished.\n- Push work in progress to a draft PR (`gh pr create --draft`) at least every 30–45 minutes;\n  the merge train never merges a draft, and the pre-push gates still run.\n- A long measurement writes each step as it finishes and resumes (`StepJournal` in\n  `docs\u002Fplans\u002Fqwenstorm-3.0.0\u002Ftools\u002Fstorm_checkpoint.py`).\n- The storm tools refuse volatile storage with exit 78 and name the key to change.\n  CONTRIBUTING.md, \"Where your work lives\", has the rest.\n\n## What each gate catches\n\n| Gate | Catches |\n|---|---|\n| ruff check | Unused imports, undefined names, syntax errors, anti-patterns |\n| ruff format | Inconsistent code style |\n| mypy --strict | Type errors, missing annotations, `Any` usage |\n| pytest (per-layer 100%) | Untested branches, logic errors, regressions |\n| lint-imports | Onion violations, forbidden imports |\n| bandit | Security issues, hardcoded secrets, insecure patterns |\n| pip-audit | Dependency CVEs |\n",{"name":5,"description":7,"allowed-tools":24},"Bash","develop","ae32736a1aabddabca5d2c96d665aa0ebc7d8700",".claude\u002Fskills\u002Fvibey-quality-gates",3,"https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey","https:\u002F\u002Fthe-vibey-project.github.io\u002Fvibey\u002F","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project","https:\u002F\u002Fvibewithadam.matthewsteinberger.com","freedom for software engineers","2026-11-15 05:47:42",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add the-vibey-project\u002Fvibey --skill vibey-quality-gates",{"tool":41,"command":42},"bunx","bunx skills add the-vibey-project\u002Fvibey --skill vibey-quality-gates",{"tool":44,"command":45},"pnpm","pnpm dlx skills add the-vibey-project\u002Fvibey --skill vibey-quality-gates","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002F.claude\u002Fskills\u002Fvibey-quality-gates\u002FSKILL.md",true,[],[50,57,64,71,78,85],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7AT1FQJAE1FQYAMDTF490Y4","white-paper-writing","Use when writing white papers, technical reports, policy documents, thought leadership pieces, or any long-form authoritative document. Triggers on requests to write, structure, outline, or review white papers, technical reports, position papers, policy briefs, or evidence-driven arguments. Also use for B2B content marketing strategy, technical documentation requiring authority, academic-style reports, or any document that must educate while persuading. Use when someone asks \"how should I structure this\" for a long-form document.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fwhite-paper-writing\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7AT16BJRJJKK2MASXRZS76G","prose-technical-mechanics","Use when editing or writing prose at a technical\u002Fmechanical level: choosing between Anglo-Saxon and Latinate diction, placing information in topic vs stress positions, writing periodic vs cumulative sentences, handling sentence rhythm, punctuation as logic, transitions, paragraph structure, or phonesthetics. Use when someone wants to understand WHY prose works or doesn't, not just surface-level style advice. Triggers on word choice, sentence structure, emphasis, rhythm, diction, or any deep writing mechanics question.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-technical-mechanics\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7AT0W9HT0ZZ9Y287NW99MR3","prose-mechanics","Use when writing, editing, or analyzing prose style; when trying to avoid generic AI-sounding writing; when applying specific prose styles such as journalistic, academic, narrative, KJV biblical, Hemingway plain, Latinate\u002FJamesian, Orwellian, Augustan, Victorian, Early Modern (Bacon, Browne, Milton), or free indirect discourse; or when coaching someone on writing voice and mechanics. Triggers on any request to improve, analyze, or write in a specific prose style, to fix \"AI-sounding\" writing, to make prose more vivid or concrete, to vary sentence rhythm, or to understand how a particular writer achieves their effects.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-mechanics\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7AT0KR635JX66X2X6F3E2CP","narrative-structure","Use when structuring stories, pitches, presentations, white papers, or any persuasive narrative; analyzing the hero's journey in fiction or non-fiction; applying StoryBrand for business communication; designing narrative arcs for thought leadership; or understanding Vogler, Snyder, Propp, or Campbell frameworks. Triggers on any narrative structure, storytelling, pitch structure, StoryBrand, hero's journey, three-act structure, Freytag, Save the Cat, narrative arc, or \"how should I structure this story\u002Fpitch\u002Fpresentation\" request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fnarrative-structure\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7AT0AKGW09NKKZDYBT8VRDE","legal-brief-writing","Use when writing legal briefs, motions, or any court filing; when analyzing judicial cognition and persuasion; when applying elite brief-writing techniques from Roberts, Clement, Kagan, Garner; or when structuring legal arguments for maximum persuasion. Also use for autistic\u002Fneurodivergent advocates who need to understand how their natural communication style maps to elite legal advocacy. Triggers on any legal writing, brief drafting, motion preparation, or legal argument structuring request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Flegal-brief-writing\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M7AT00VWPS4Q8B5VEMXX9NQT","journal-publishability-criteria","Use whenever a research paper, article or preprint is being written, revised, reviewed or published, and specifically before every publication of docs\u002Fpaper.md. Covers what legitimate journals actually gate on, in order of gating — the editor's desk screen (scope, novelty, package problems), what reviewers judge (methodological soundness, claims the evidence can carry, statistical and qualitative reporting), EQUATOR reporting guidelines, the ICMJE authorship criteria, COPE's view of text recycling, the near-universal generative-AI policy (no AI author, disclosure, author responsibility), TOP 2025 transparency standards, how to verify a journal is legitimate, submitting without an institution, and a pre-submission checklist. Triggers on requests to evaluate, review, prepare or submit a paper, to judge publishability, to write a declarations or data-availability section, to disclose AI use, or to check a journal.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fjournal-publishability-criteria\u002FSKILL.md",[],[]]