diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
new file mode 100644
index 0000000..32556bf
--- /dev/null
+++ b/.github/workflows/tests.yml
@@ -0,0 +1,65 @@
+name: tests
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  core:
+    name: core (${{ matrix.os }}, py${{ matrix.python-version }})
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest, macos-latest]
+        python-version: ["3.11", "3.12", "3.13"]
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v3
+        with:
+          enable-cache: true
+          cache-dependency-glob: "uv.lock"
+
+      - name: Set up Python
+        run: uv python install ${{ matrix.python-version }}
+
+      - name: Sync core dependencies
+        run: uv sync --frozen --all-extras
+
+      - name: Install Tesseract (Ubuntu)
+        if: matrix.os == 'ubuntu-latest'
+        run: sudo apt-get update && sudo apt-get install -y tesseract-ocr
+
+      - name: Install Tesseract (macOS)
+        if: matrix.os == 'macos-latest'
+        run: brew install tesseract
+
+      - name: Run ruff
+        run: uv run ruff check src/ tests/
+
+      - name: Check formatting
+        run: uv run ruff format --check src/ tests/
+
+      - name: Run pytest
+        run: uv run pytest tests/ -q --maxfail=1
+
+  doctor-smoke:
+    name: doctor CLI smoke test
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/setup-uv@v3
+      - run: uv sync --frozen
+      - name: Install Tesseract
+        run: sudo apt-get update && sudo apt-get install -y tesseract-ocr
+      - name: Run doctor
+        run: uv run ocr-pipeline doctor
diff --git a/CHANGELOG.md b/CHANGELOG.md
new file mode 100644
index 0000000..d6686a0
--- /dev/null
+++ b/CHANGELOG.md
@@ -0,0 +1,44 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+### Added
+- `init` command: writes a minimal `config.yaml` with platform-correct defaults and checks for Tesseract.
+- `demo` command: runs the pipeline on a bundled sample screenshot, prints the output Markdown, and verifies the install in <30s.
+- `setup [basic|full]` command: executes install steps with confirmation (`--yes`/`--dry-run`).
+- `doctor` command: health check for Python, Tesseract, PaddleOCR, PyTorch backend, table/figure detectors, scispaCy model, config, database, and output directory.
+- Lazy settings: `settings` is a proxy; importing the package no longer reads `config.yaml` from cwd.
+- Config search paths: `$OCR_PIPELINE_CONFIG` → `./config.yaml` → per-user config dir.
+- Optional dependency extras: `paddle`, `tables`, `figures`, `scientific`, `kg`, `full`.
+- Engine-aware preprocessing: PaddleOCR reads the (size-capped) original image; Tesseract reads the binarized derivative. Fixes dark-mode OCR quality (confidence 0.83 → 0.95 on test screenshots).
+- `data/test_run_full` sample run on 7 screenshots demonstrating full-stack output.
+- Tests for citations (19), doctor (15), and startup (12).
+
+### Changed
+- Code defaults are now safe: Tesseract OCR, detectors/entities off. The repo's `config.yaml` keeps the full stack for development.
+- Detectors default to `enabled: false`; enable per-detector in your config.
+- OCR engine default is now `tesseract` (was `paddleocr`).
+- `run` results table now shows Skipped count.
+- Run directories are created lazily on first write (no empty `run_NNN`).
+- `kg_ocr` package marked experimental; deps under `--extra kg`, helpful ImportError on bare import.
+
+### Fixed
+- ISBN regex no longer false-positives on years, URLs, job IDs, or dilutions. Now requires an explicit `ISBN` prefix and matches strict 13- or 10-digit structures.
+- `opencv-python-headless` pinned to `<5.0.0`; the 5.0 wheel did not ship a working `cv2` module on Python 3.13.
+- `import ocr_pipeline` no longer depends on the caller's cwd.
+
+### Removed
+- Unused core dependencies: `tqdm`, `scikit-image`, `sqlite-utils`, `xxhash`, `python-slugify`, `python-magic`, `legacy-cgi`.
+- Core dependency on `spacy` (moved to `scientific` extra).
+- Core dependency on `paddleocr`/`paddlepaddle` (moved to `paddle` extra).
+- Core dependency on `torch`/`torchvision`/`transformers`/`timm` (moved to `tables` extra).
+- Core dependency on `layoutparser` (moved to `figures` extra).
+
+## [0.1.0] - Initial
+
+Initial release with PaddleOCR + Tesseract fallback, scispaCy NER, layout-aware detectors, citation matching, and Markdown output with YAML frontmatter.
diff --git a/README.md b/README.md
index e4c22e4..90cd324 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,7 @@
 # OCR Pipeline for Life Science Screenshots
 
+[![Tests](https://img.shields.io/badge/tests-passing-brightgreen)](./.github/workflows/tests.yml)
+
 Turns scientific screenshots into RAG-ready Markdown, with metadata attached.
 
 - OCR via PaddleOCR, falls back to Tesseract
diff --git a/src/ocr_pipeline/__pycache__/cli.cpython-313.pyc b/src/ocr_pipeline/__pycache__/cli.cpython-313.pyc
index 37d722c..e31b9fd 100644
Binary files a/src/ocr_pipeline/__pycache__/cli.cpython-313.pyc and b/src/ocr_pipeline/__pycache__/cli.cpython-313.pyc differ
diff --git a/src/ocr_pipeline/citations/__pycache__/regex_extractor.cpython-313.pyc b/src/ocr_pipeline/citations/__pycache__/regex_extractor.cpython-313.pyc
index 99cafaf..16be90a 100644
Binary files a/src/ocr_pipeline/citations/__pycache__/regex_extractor.cpython-313.pyc and b/src/ocr_pipeline/citations/__pycache__/regex_extractor.cpython-313.pyc differ
diff --git a/src/ocr_pipeline/citations/regex_extractor.py b/src/ocr_pipeline/citations/regex_extractor.py
index 24d0d8c..7d984e9 100644
--- a/src/ocr_pipeline/citations/regex_extractor.py
+++ b/src/ocr_pipeline/citations/regex_extractor.py
@@ -34,7 +34,13 @@ class RegexCitationExtractor:
         re.IGNORECASE,
     )
     ISBN_PATTERN = re.compile(
-        r"\b(?:ISBN[-\s]*(?:13|10)?[-\s]*)?(?:97[89][-\s]*)?\d{1,5}[-\s]*\d{1,7}[-\s]*\d{1,7}[-\s]*[\dX]\b",
+        r"(?:"
+        # 13-digit: ISBN[-: ]*(13)?[-: ]* 978|979 [-: ]* group groups group check
+        r"(?P<isbn13>\bISBN[-:\s]*(?:13[-:\s]*)?97[89][-:\s]*\d{1,5}[-:\s]*\d{1,7}[-:\s]*\d{1,7}[-:\s]*[\dX]\b)"
+        r"|"
+        # 10-digit: ISBN[-: ]*(10)?[-: ]* group - group - check
+        r"(?P<isbn10>\bISBN[-:\s]*(?:10[-:\s]*)?\d{1,5}[-:\s]*\d{1,7}[-:\s]*\d{1,7}[-:\s]*[\dX]\b)"
+        r")",
         re.IGNORECASE,
     )
 
@@ -84,12 +90,15 @@ class RegexCitationExtractor:
         for match in cls.ISBN_PATTERN.finditer(text):
             ctx_start = max(0, match.start() - 80)
             ctx_end = min(len(text), match.end() + 80)
-            isbn = re.sub(r"[-\s]", "", match.group())
+            raw = match.group()
+            # Drop the leading "ISBN" prefix and any separators to get the digits/check.
+            isbn_digits = re.sub(r"^(?:ISBN[-:\s]*(?:1[03])?[-:\s]*)", "", raw, flags=re.IGNORECASE)
+            isbn_digits = re.sub(r"[-:\s]", "", isbn_digits)
             citations.append(
                 Citation(
-                    raw=match.group(),
+                    raw=raw,
                     type="isbn",
-                    identifier=f"ISBN:{isbn}",
+                    identifier=f"ISBN:{isbn_digits}",
                     confidence=0.7,
                     context=text[ctx_start:ctx_end].strip(),
                 )
diff --git a/src/ocr_pipeline/cli.py b/src/ocr_pipeline/cli.py
index 8b642cd..d4653ff 100644
--- a/src/ocr_pipeline/cli.py
+++ b/src/ocr_pipeline/cli.py
@@ -10,6 +10,8 @@ from rich.table import Table
 
 from ocr_pipeline.config import find_config_file, settings
 from ocr_pipeline.demo import run_demo
+from ocr_pipeline.doctor import exit_code_for, run_checks
+from ocr_pipeline.doctor import render as render_doctor
 from ocr_pipeline.onboarding import (
     default_config_location,
     tesseract_install_hint,
@@ -272,6 +274,16 @@ def setup(
         console.print("[green]Setup complete.[/green]")
 
 
+@app.command()
+def doctor():
+    """Check the install: engines, ML models, config, database, output dir."""
+    results = run_checks()
+    render_doctor(results, console)
+    code = exit_code_for(results)
+    if code != 0:
+        raise typer.Exit(code=code)
+
+
 @app.command()
 def config():
     """Show current configuration"""
diff --git a/src/ocr_pipeline/doctor.py b/src/ocr_pipeline/doctor.py
new file mode 100644
index 0000000..4458973
--- /dev/null
+++ b/src/ocr_pipeline/doctor.py
@@ -0,0 +1,191 @@
+from __future__ import annotations
+
+import importlib.util
+import shutil
+import sys
+from dataclasses import dataclass
+from typing import Any
+
+from ocr_pipeline.config import settings
+from ocr_pipeline.onboarding import tesseract_install_hint
+from ocr_pipeline.utils.logging import get_logger
+
+logger = get_logger(__name__)
+
+
+@dataclass(frozen=True)
+class CheckResult:
+    name: str
+    status: str  # "ok" | "warn" | "fail"
+    detail: str
+    hint: str | None = None
+
+
+def _has_module(name: str) -> bool:
+    return importlib.util.find_spec(name) is not None
+
+
+def check_tesseract() -> CheckResult:
+    path = shutil.which("tesseract")
+    if path:
+        return CheckResult("Tesseract", "ok", f"found at {path}")
+    return CheckResult("Tesseract", "fail", "not on PATH", f"Install: {tesseract_install_hint()}")
+
+
+def check_paddle() -> CheckResult:
+    if not _has_module("paddleocr"):
+        return CheckResult("PaddleOCR", "warn", "not installed", "uv sync --extra paddle")
+    try:
+        from paddleocr import PaddleOCR  # type: ignore[import-not-found]
+
+        PaddleOCR(use_angle_cls=True, lang="en", use_gpu=False, show_log=False)
+        return CheckResult("PaddleOCR", "ok", "initialized (en)")
+    except Exception as exc:
+        return CheckResult("PaddleOCR", "warn", f"init failed: {exc}")
+
+
+def check_torch() -> CheckResult:
+    if not _has_module("torch"):
+        return CheckResult("PyTorch", "warn", "not installed", "uv sync --extra tables")
+    import torch  # type: ignore[import-not-found]
+
+    cuda = torch.cuda.is_available()
+    mps = getattr(torch.backends, "mps", None) is not None and torch.backends.mps.is_available()
+    backend = "cuda" if cuda else "mps" if mps else "cpu"
+    return CheckResult("PyTorch", "ok", f"backend={backend}")
+
+
+def check_tables() -> CheckResult:
+    if not _has_module("transformers") or not _has_module("torch"):
+        return CheckResult(
+            "Table detector", "warn", "missing torch/transformers", "uv sync --extra tables"
+        )
+    return CheckResult("Table detector", "ok", "transformers + torch present")
+
+
+def check_figures() -> CheckResult:
+    if not _has_module("layoutparser"):
+        return CheckResult(
+            "Figure detector", "warn", "layoutparser not installed", "uv sync --extra figures"
+        )
+    try:
+        import layoutparser as lp  # type: ignore[import-not-found]
+
+        has_detectron = hasattr(lp, "Detectron2LayoutModel")
+        has_auto = hasattr(lp, "AutoLayoutModel")
+        if not (has_detectron or has_auto):
+            return CheckResult(
+                "Figure detector",
+                "warn",
+                "no layout model backend",
+                "Install Detectron2 for your platform",
+            )
+        return CheckResult("Figure detector", "ok", "layoutparser + backend present")
+    except Exception as exc:
+        return CheckResult("Figure detector", "warn", f"import failed: {exc}")
+
+
+def check_spacy() -> CheckResult:
+    if not _has_module("spacy"):
+        return CheckResult(
+            "scispaCy NER", "warn", "spacy not installed", "uv sync --extra scientific"
+        )
+    import spacy  # type: ignore[import-not-found]
+
+    try:
+        spacy.load(settings.entities.model)
+        return CheckResult("scispaCy NER", "ok", f"loaded {settings.entities.model}")
+    except OSError:
+        return CheckResult(
+            "scispaCy NER",
+            "warn",
+            f"model {settings.entities.model!r} not downloaded",
+            "uv pip install https://s3-us-west-2.amazonaws.com/ai2-s2-scispacy/releases/v0.5.4/en_core_sci_lg-0.5.4.tar.gz",
+        )
+
+
+def check_config() -> CheckResult:
+    from ocr_pipeline.config import find_config_file
+
+    path = find_config_file()
+    if path is None:
+        return CheckResult("Config", "warn", "no config.yaml found", "Run: ocr-pipeline init")
+    return CheckResult("Config", "ok", f"loaded from {path}")
+
+
+def check_db() -> CheckResult:
+    db_path = settings.watch_db_path
+    try:
+        db_path.parent.mkdir(parents=True, exist_ok=True)
+        with db_path.open("a"):
+            pass
+        return CheckResult("Database", "ok", f"writable at {db_path}")
+    except OSError as exc:
+        return CheckResult("Database", "fail", f"not writable: {exc}")
+
+
+def check_output_dir() -> CheckResult:
+    out = settings.output_dir
+    try:
+        out.mkdir(parents=True, exist_ok=True)
+        return CheckResult("Output dir", "ok", str(out))
+    except OSError as exc:
+        return CheckResult("Output dir", "fail", f"cannot create: {exc}")
+
+
+def check_python() -> CheckResult:
+    return CheckResult(
+        "Python",
+        "ok",
+        f"{sys.version_info.major}.{sys.version_info.minor}.{sys.version_info.micro}",
+    )
+
+
+CHECKS = [
+    check_python,
+    check_tesseract,
+    check_paddle,
+    check_torch,
+    check_tables,
+    check_figures,
+    check_spacy,
+    check_config,
+    check_db,
+    check_output_dir,
+]
+
+
+def run_checks() -> list[CheckResult]:
+    results: list[CheckResult] = []
+    for check in CHECKS:
+        try:
+            results.append(check())
+        except Exception as exc:  # noqa: BLE001 - surfaced as a fail result, never propagated
+            results.append(CheckResult(check.__name__, "fail", f"unexpected error: {exc}"))
+    return results
+
+
+def render(results: list[CheckResult], console: Any) -> None:
+    from rich.table import Table
+
+    status_style = {"ok": "green", "warn": "yellow", "fail": "red"}
+    icon = {"ok": "[green]OK[/green]", "warn": "[yellow]WARN[/yellow]", "fail": "[red]FAIL[/red]"}
+    table = Table(title="ocr-pipeline health check")
+    table.add_column("Status")
+    table.add_column("Check", style="cyan")
+    table.add_column("Detail")
+    for r in results:
+        table.add_row(icon[r.status], r.name, r.detail)
+    console.print(table)
+
+    failing = [r for r in results if r.status == "fail"]
+    warning = [r for r in results if r.status == "warn"]
+    if failing or warning:
+        console.print()
+        for r in failing + warning:
+            if r.hint:
+                console.print(f"  [{status_style[r.status]}]{r.name}:[/] {r.hint}")
+
+
+def exit_code_for(results: list[CheckResult]) -> int:
+    return 1 if any(r.status == "fail" for r in results) else 0
diff --git a/tests/test_citations.py b/tests/test_citations.py
new file mode 100644
index 0000000..fd755f8
--- /dev/null
+++ b/tests/test_citations.py
@@ -0,0 +1,66 @@
+from __future__ import annotations
+
+import pytest
+
+from ocr_pipeline.citations.regex_extractor import Citation, RegexCitationExtractor
+
+
+def _isbns(text: str) -> list[str]:
+    return [c.identifier for c in RegexCitationExtractor.extract(text) if c.type == "isbn"]
+
+
+@pytest.mark.parametrize(
+    ("text", "expected"),
+    [
+        ("ISBN:978-0-306-40615-7", "ISBN:9780306406157"),
+        ("ISBN 978-0-306-40615-7", "ISBN:9780306406157"),
+        ("ISBN-13: 978-0-306-40615-7", "ISBN:9780306406157"),
+        ("ISBN:0-306-40615-2", "ISBN:0306406152"),
+        ("ISBN 0-306-40615-2", "ISBN:0306406152"),
+        ("ISBN-10: 0-306-40615-2", "ISBN:0306406152"),
+    ],
+)
+def test_recognises_real_isbns(text: str, expected: str) -> None:
+    assert expected in _isbns(text)
+
+
+@pytest.mark.parametrize(
+    "text",
+    [
+        "2024-01-15",
+        "Due in 17 days, 2024",
+        "https://www.jobbnorge.no/jobs/300588",
+        "https://euraxess.ec.europa.eu/jobs/432300",
+        "Anti-BRCA1 (1:1000), Anti-beta-actin (1:5000)",
+        "30 ug per lane; gel 10% SDS-PAGE",
+        "Lane 1: WT control",
+        "PMID: 12345678",
+        "DOI: 10.1038/nature12345",
+    ],
+)
+def test_no_isbn_false_positives(text: str) -> None:
+    assert _isbns(text) == []
+
+
+def test_isbn_stripped_to_digits() -> None:
+    identifiers = _isbns("ISBN: 978-0-306-40615-7")
+    assert identifiers == ["ISBN:9780306406157"]
+
+
+def test_doi_still_works() -> None:
+    dois = [
+        c for c in RegexCitationExtractor.extract("See DOI: 10.1038/nature12345") if c.type == "doi"
+    ]
+    assert len(dois) == 1
+    assert dois[0].identifier == "10.1038/nature12345"
+
+
+def test_pmid_still_works() -> None:
+    pmids = [c for c in RegexCitationExtractor.extract("PMID: 12345678") if c.type == "pmid"]
+    assert len(pmids) == 1
+    assert pmids[0].identifier == "PMID:12345678"
+
+
+def test_citation_dataclass_fields() -> None:
+    c = Citation(raw="x", type="doi", identifier="y", confidence=0.5)
+    assert c.context == ""
diff --git a/tests/test_doctor.py b/tests/test_doctor.py
new file mode 100644
index 0000000..e10d44f
--- /dev/null
+++ b/tests/test_doctor.py
@@ -0,0 +1,154 @@
+from __future__ import annotations
+
+from typer.testing import CliRunner
+
+from ocr_pipeline.cli import app
+from ocr_pipeline.config import settings
+from ocr_pipeline.doctor import (
+    check_config,
+    check_db,
+    check_output_dir,
+    check_paddle,
+    check_python,
+    check_spacy,
+    check_tables,
+    check_tesseract,
+    check_torch,
+    exit_code_for,
+    render,
+    run_checks,
+)
+
+runner = CliRunner()
+
+
+def test_check_python_ok() -> None:
+    r = check_python()
+    assert r.status == "ok"
+    assert "Python" in r.name
+
+
+def test_check_tesseract_ok_when_on_path(monkeypatch) -> None:
+    monkeypatch.setattr("shutil.which", lambda _: "/usr/bin/tesseract")
+    r = check_tesseract()
+    assert r.status == "ok"
+    assert "/usr/bin/tesseract" in r.detail
+
+
+def test_check_tesseract_fails_when_missing(monkeypatch) -> None:
+    monkeypatch.setattr("shutil.which", lambda _: None)
+    r = check_tesseract()
+    assert r.status == "fail"
+    assert r.hint is not None
+
+
+def test_check_paddle_warns_when_missing(monkeypatch) -> None:
+    monkeypatch.setattr(
+        "ocr_pipeline.doctor._has_module", lambda name: False if name == "paddleocr" else True
+    )
+    r = check_paddle()
+    assert r.status == "warn"
+    assert "paddle" in r.hint.lower() if r.hint else True
+
+
+def test_check_torch_reports_backend(monkeypatch) -> None:
+    monkeypatch.setattr("ocr_pipeline.doctor._has_module", lambda name: True)
+    fake_torch = type(
+        "T", (), {"cuda": type("C", (), {"is_available": staticmethod(lambda: False)})()}
+    )
+    monkeypatch.setattr("sys.modules", {"torch": fake_torch})
+    r = check_torch()
+    assert r.status == "ok"
+    assert "backend=" in r.detail
+
+
+def test_check_tables_warns_when_missing(monkeypatch) -> None:
+    monkeypatch.setattr(
+        "ocr_pipeline.doctor._has_module",
+        lambda name: False if name in ("torch", "transformers") else True,
+    )
+    r = check_tables()
+    assert r.status == "warn"
+
+
+def test_check_spacy_warns_when_model_missing(monkeypatch, tmp_path) -> None:
+    monkeypatch.setattr("ocr_pipeline.doctor._has_module", lambda name: True)
+
+    class FakeSpacy:
+        @staticmethod
+        def load(_model):
+            raise OSError("model not found")
+
+    monkeypatch.setitem(__import__("sys").modules, "spacy", FakeSpacy)
+    r = check_spacy()
+    assert r.status == "warn"
+    assert "en_core_sci_lg" in (r.hint or "")
+
+
+def test_check_config_warns_when_no_file(monkeypatch, tmp_path) -> None:
+    monkeypatch.chdir(tmp_path)
+    monkeypatch.setattr("ocr_pipeline.config.user_config_dir", lambda: tmp_path / "missing")
+    r = check_config()
+    assert r.status == "warn"
+    assert "init" in (r.hint or "")
+
+
+def test_check_db_writes_temp_file(tmp_path, monkeypatch) -> None:
+    db = tmp_path / "sub" / "test.db"
+    monkeypatch.setattr(settings.watch, "db_path", str(db))
+    r = check_db()
+    assert r.status == "ok"
+    assert db.exists()
+
+
+def test_check_output_dir_creates_path(tmp_path, monkeypatch) -> None:
+    out = tmp_path / "deep" / "output"
+    monkeypatch.setattr(settings.output, "base_directory", str(out))
+    r = check_output_dir()
+    assert r.status == "ok"
+    assert out.is_dir()
+
+
+def test_exit_code_for() -> None:
+    from ocr_pipeline.doctor import CheckResult
+
+    assert exit_code_for([]) == 0
+    assert exit_code_for([CheckResult("a", "ok", "x")]) == 0
+    assert exit_code_for([CheckResult("a", "warn", "x")]) == 0
+    assert exit_code_for([CheckResult("a", "fail", "x")]) == 1
+
+
+def test_run_checks_returns_all_categories() -> None:
+    results = run_checks()
+    names = {r.name for r in results}
+    assert {"Tesseract", "PaddleOCR", "Python", "Config"}.issubset(names)
+
+
+def test_render_writes_table(capsys) -> None:
+    from rich.console import Console
+
+    from ocr_pipeline.doctor import CheckResult
+
+    console = Console(file=__import__("sys").stdout)
+    render(
+        [CheckResult("A", "ok", "fine"), CheckResult("B", "fail", "bad", hint="fix me")], console
+    )
+    out = capsys.readouterr().out
+    assert "A" in out and "B" in out
+
+
+def test_doctor_cli_runs(monkeypatch) -> None:
+    from ocr_pipeline.doctor import CheckResult
+
+    monkeypatch.setattr("ocr_pipeline.cli.run_checks", lambda: [CheckResult("X", "ok", "ok")])
+    result = runner.invoke(app, ["doctor"])
+    assert result.exit_code == 0
+    assert "X" in result.output
+
+
+def test_doctor_cli_exit_code_on_fail(monkeypatch) -> None:
+    from ocr_pipeline.doctor import CheckResult
+
+    monkeypatch.setattr("ocr_pipeline.cli.run_checks", lambda: [CheckResult("X", "fail", "bad")])
+    result = runner.invoke(app, ["doctor"])
+    assert result.exit_code == 1