diff --git a/in2lambda/validation/__init__.py b/in2lambda/validation/__init__.py
new file mode 100644
index 0000000..b4400c0
--- /dev/null
+++ b/in2lambda/validation/__init__.py
@@ -0,0 +1,40 @@
+"""Pre-flight checks for the ``#``/``##`` markdown that flows through in2lambda.
+
+The markdown produced by the wizard (and hand-written by users) is the shared
+contract between the wizard, the ``Markdown`` filter and Lambda Feedback. These
+checks catch structural mistakes - currently unbalanced/misplaced math
+delimiters - before the markdown is converted.
+"""
+
+from in2lambda.validation.delimiters import (
+    MathDelimiterError,
+    MathDelimiterProblem,
+    math_delimiter_checker,
+)
+
+__all__ = [
+    "MathDelimiterError",
+    "MathDelimiterProblem",
+    "math_delimiter_checker",
+    "check_markdown",
+]
+
+
+def check_markdown(md_content: str) -> list[MathDelimiterProblem]:
+    """Run every markdown check and return the problems found.
+
+    Args:
+        md_content: The markdown text to validate.
+
+    Returns:
+        A list of :class:`MathDelimiterProblem`, one per problem found.
+        An empty list means the markdown passed every check.
+
+    Examples:
+        >>> from in2lambda.validation import check_markdown
+        >>> check_markdown("Inline $x = y$ is fine.")
+        []
+        >>> check_markdown("Unbalanced $x = y")
+        [MathDelimiterProblem(line=1, error=<MathDelimiterError.MISSING_CLOSING_SINGLE_DOLLAR: 'unclosed inline $ ... $'>)]
+    """
+    return math_delimiter_checker(md_content)
diff --git a/in2lambda/validation/delimiters.py b/in2lambda/validation/delimiters.py
new file mode 100644
index 0000000..10b8014
--- /dev/null
+++ b/in2lambda/validation/delimiters.py
@@ -0,0 +1,199 @@
+"""Checks that ``$ ... $`` and ``$$ ... $$`` math delimiters are balanced and placed correctly.
+
+Lambda Feedback renders inline math (via KaTeX) wrapped in single dollar
+signs, and display math wrapped in ``$$``. Requiring each ``$$`` delimiter to
+sit alone on its own line is this project's own authoring convention, not a
+KaTeX requirement. This module scans markdown character by character,
+skipping fenced and inline code, and reports every delimiter mistake it
+finds together with the line it occurred on.
+"""
+
+from dataclasses import dataclass
+from enum import Enum
+
+
+class MathDelimiterError(Enum):
+    """A specific delimiter mistake found by :func:`math_delimiter_checker`.
+
+    The value is a short human-readable message suitable for showing on the
+    command line.
+    """
+
+    MISSING_NEWLINE_BEFORE_OPENING_DISPLAY = "opening $$ must start its own line"
+    MISSING_NEWLINE_AFTER_OPENING_DISPLAY = "opening $$ must be followed by a newline"
+    DOUBLE_DOLLAR_INSTEAD_OF_CLOSING_SINGLE = "inline $ ... $ closed with $$"
+    MISSING_CLOSING_DOUBLE_INSTEAD_OF_SINGLE = (
+        "display $$ ... $$ closed with a single $"
+    )
+    MISSING_NEWLINE_BEFORE_CLOSING_DISPLAY = "closing $$ must start its own line"
+    MISSING_NEWLINE_AFTER_CLOSING_DISPLAY = "closing $$ must be followed by a newline"
+    MISSING_CLOSING_SINGLE_DOLLAR = "unclosed inline $ ... $"
+    MISSING_CLOSING_DOUBLE_DOLLAR = "unclosed display $$ ... $$"
+
+
+@dataclass(frozen=True)
+class MathDelimiterProblem:
+    """A single delimiter mistake and the (1-based) line it was found on."""
+
+    line: int
+    error: MathDelimiterError
+
+    def __str__(self) -> str:
+        """Render as ``line <n>: <message>``."""
+        return f"line {self.line}: {self.error.value}"
+
+
+def math_delimiter_checker(md_content: str) -> list[MathDelimiterProblem]:
+    r"""Scan markdown for every math-delimiter mistake.
+
+    ``\$`` is treated as a literal dollar sign, not a delimiter.
+
+    Args:
+        md_content: The markdown text to check.
+
+    Returns:
+        A list of :class:`MathDelimiterProblem`, one per mistake found, in
+        the order they occur. An empty list means the delimiters are well
+        formed.
+
+    Examples:
+        >>> from in2lambda.validation.delimiters import math_delimiter_checker
+        >>> math_delimiter_checker("An inline $x = y$ expression.")
+        []
+        >>> math_delimiter_checker("Display:\n$$\nx = y\n$$")
+        []
+        >>> math_delimiter_checker("This costs \\$5, no math here.")
+        []
+        >>> math_delimiter_checker("Run `echo $PATH` now.")
+        []
+        >>> math_delimiter_checker("Broken $x = y")
+        [MathDelimiterProblem(line=1, error=<MathDelimiterError.MISSING_CLOSING_SINGLE_DOLLAR: 'unclosed inline $ ... $'>)]
+    """
+    problems: list[MathDelimiterProblem] = []
+
+    def report(error: MathDelimiterError) -> None:
+        problems.append(MathDelimiterProblem(md_content.count("\n", 0, idx) + 1, error))
+
+    # False once we are inside a math expression and awaiting its closing delimiter.
+    expect_open_delimiter = True
+    # While inside an expression, whether it opened with a single "$" (inline) or "$$" (display).
+    expect_single_dollar = True
+    # Line on which the currently open (unclosed) expression started.
+    open_line = 1
+
+    # Backtick code spans/fences are not markdown math and must not be scanned for
+    # "$" delimiters, e.g. a shell variable like `echo $PATH` or a fenced snippet.
+    in_fence = False
+    fence_marker_len = 0
+    in_code_span = False
+    code_span_marker_len = 0
+
+    idx = 0
+    while idx < len(md_content):
+        prev_character = md_content[idx - 1] if idx > 0 else None
+        character = md_content[idx]
+        next_character = md_content[idx + 1] if idx + 1 < len(md_content) else None
+
+        if character == "`" and prev_character != "`":
+            run_len = 0
+            while idx + run_len < len(md_content) and md_content[idx + run_len] == "`":
+                run_len += 1
+            line_start = md_content.rfind("\n", 0, idx) + 1
+            at_line_start = md_content[line_start:idx].strip() == ""
+
+            if in_fence:
+                line_end = md_content.find("\n", idx + run_len)
+                if line_end == -1:
+                    line_end = len(md_content)
+                if (
+                    run_len >= fence_marker_len
+                    and at_line_start
+                    and md_content[idx + run_len : line_end].strip() == ""
+                ):
+                    in_fence, fence_marker_len = False, 0
+            elif in_code_span:
+                if run_len == code_span_marker_len:
+                    in_code_span, code_span_marker_len = False, 0
+            elif run_len >= 3 and at_line_start:
+                in_fence, fence_marker_len = True, run_len
+            else:
+                in_code_span, code_span_marker_len = True, run_len
+
+            idx += 1
+            continue
+
+        if in_fence or in_code_span:
+            idx += 1
+            continue
+
+        if character == "$" and prev_character != "\\":
+            if expect_open_delimiter:
+                expect_open_delimiter = False
+                open_line = md_content.count("\n", 0, idx) + 1
+
+                if next_character == "$":
+                    next_next_character = (
+                        md_content[idx + 2] if idx + 2 < len(md_content) else None
+                    )
+                    # "$$" must sit alone on its own line.
+                    if prev_character != "\n" and prev_character is not None:
+                        report(
+                            MathDelimiterError.MISSING_NEWLINE_BEFORE_OPENING_DISPLAY
+                        )
+                        expect_open_delimiter, expect_single_dollar = True, True
+                    elif next_next_character != "\n":
+                        report(MathDelimiterError.MISSING_NEWLINE_AFTER_OPENING_DISPLAY)
+                        expect_open_delimiter, expect_single_dollar = True, True
+                    else:
+                        expect_single_dollar = False
+                    idx += 1  # Skip the second "$"; the loop increments idx again.
+                else:
+                    expect_single_dollar = True
+            else:
+                expect_open_delimiter = True
+
+                if expect_single_dollar and next_character == "$":
+                    report(MathDelimiterError.DOUBLE_DOLLAR_INSTEAD_OF_CLOSING_SINGLE)
+                    expect_open_delimiter, expect_single_dollar = True, True
+
+                elif not expect_single_dollar:
+                    if next_character != "$":
+                        report(
+                            MathDelimiterError.MISSING_CLOSING_DOUBLE_INSTEAD_OF_SINGLE
+                        )
+                        expect_open_delimiter, expect_single_dollar = True, True
+                    else:
+                        next_next_character = (
+                            md_content[idx + 2] if idx + 2 < len(md_content) else None
+                        )
+                        if prev_character != "\n" and prev_character is not None:
+                            report(
+                                MathDelimiterError.MISSING_NEWLINE_BEFORE_CLOSING_DISPLAY
+                            )
+                            expect_open_delimiter, expect_single_dollar = True, True
+                        elif (
+                            next_next_character != "\n"
+                            and next_next_character is not None
+                        ):
+                            report(
+                                MathDelimiterError.MISSING_NEWLINE_AFTER_CLOSING_DISPLAY
+                            )
+                            expect_open_delimiter, expect_single_dollar = True, True
+
+                        idx += 1  # Skip the second "$"; the loop increments idx again.
+
+        idx += 1
+
+    if not expect_open_delimiter:
+        problems.append(
+            MathDelimiterProblem(
+                open_line,
+                (
+                    MathDelimiterError.MISSING_CLOSING_SINGLE_DOLLAR
+                    if expect_single_dollar
+                    else MathDelimiterError.MISSING_CLOSING_DOUBLE_DOLLAR
+                ),
+            )
+        )
+
+    return problems
diff --git a/tests/test_validation.py b/tests/test_validation.py
new file mode 100644
index 0000000..698309e
--- /dev/null
+++ b/tests/test_validation.py
@@ -0,0 +1,157 @@
+"""Tests for the math-delimiter checker.
+
+Ported from ``conversion2025/tools and testing/validator_tests.py`` on the
+``Summer2025`` branch and adapted to the :class:`MathDelimiterError` enum.
+"""
+
+import pytest
+
+from in2lambda.validation import (
+    MathDelimiterError,
+    MathDelimiterProblem,
+    check_markdown,
+    math_delimiter_checker,
+)
+
+E = MathDelimiterError
+P = MathDelimiterProblem
+
+VALID = [
+    "This is an inline math expression: $x = y$.",
+    "This is an inline math expression: $x = y$",
+    "$x = y$, this is an inline math expression.",
+    "$x = y$\n",
+    "\n$x = y$",
+    "First expression $x = y$ and second expression $a = b$.",
+    "Expression: $\\alpha + \\beta = \\gamma$.",
+    "This is a display math expression:\n$$\nx = y\n$$",
+    "$$\nx = y\n$$\n, this is a display math expression.",
+    "Display math:\n$$\nx = y\n\na = b\n$$",
+    "Expression:\n$$\n$$",
+    "Inline $x = y$ and display math:\n$$\nx = y\n$$",
+    "First:\n$$\nx = y\n$$\nSecond:\n$$\na = b\n$$",
+    "",
+    "This is just regular text with no math expressions.",
+    "This costs \\$5 and that costs \\$10.",
+    "Price is \\$10 and math is $x = y$.",
+    "Price \\$100:\n$$\nx = y\n$$",
+    "This symbol \\$\\$ is not math.",
+    "\\$100 is expensive.",
+    "It costs \\$",
+    "Price \\$50 for $x + y = z$ calculation.",
+    "Expression: $cost = \\$100$.",
+    "Display:\n$$\ncost = \\$100\n$$",
+    # A "$ ... $" expression may now span a newline before closing.
+    "This is an inline math expression:$x \n= y$.",
+    "This is an inline math expression:$\nx = y$.",
+    "This is an inline math expression:$x = y\n$.",
+    # A "$" inside a code span/fence is not math and must be ignored.
+    "Run `echo $PATH` now.\n",
+    "```bash\necho $HOME\n```\n",
+]
+
+INVALID = [
+    (
+        "This is an inline math expression: $x = y.",
+        [P(1, E.MISSING_CLOSING_SINGLE_DOLLAR)],
+    ),
+    (
+        "This is an inline math expression: x = y$.",
+        [P(1, E.MISSING_CLOSING_SINGLE_DOLLAR)],
+    ),
+    (
+        "Expression $x = y$$.",
+        [
+            P(1, E.DOUBLE_DOLLAR_INSTEAD_OF_CLOSING_SINGLE),
+            P(1, E.MISSING_CLOSING_SINGLE_DOLLAR),
+        ],
+    ),
+    (
+        "Expression $x = y$ and $a = b$ and $c =",
+        [P(1, E.MISSING_CLOSING_SINGLE_DOLLAR)],
+    ),
+    (
+        "Expression $$$x = y$$$.",
+        [
+            P(1, E.MISSING_NEWLINE_BEFORE_OPENING_DISPLAY),
+            P(1, E.DOUBLE_DOLLAR_INSTEAD_OF_CLOSING_SINGLE),
+            P(1, E.MISSING_NEWLINE_BEFORE_OPENING_DISPLAY),
+        ],
+    ),
+    (
+        "This is a display math expression:\n$$\nx = y\n",
+        [P(2, E.MISSING_CLOSING_DOUBLE_DOLLAR)],
+    ),
+    (
+        "This is a display math expression:\nx = y\n$$",
+        [P(3, E.MISSING_NEWLINE_AFTER_OPENING_DISPLAY)],
+    ),
+    (
+        "This is a display math expression:$$\nx = y\n$$.",
+        [
+            P(1, E.MISSING_NEWLINE_BEFORE_OPENING_DISPLAY),
+            P(3, E.MISSING_NEWLINE_AFTER_OPENING_DISPLAY),
+        ],
+    ),
+    (
+        "This is a display math expression:\n$$\nx = y\n$$.",
+        [P(4, E.MISSING_NEWLINE_AFTER_CLOSING_DISPLAY)],
+    ),
+    (
+        "Expression:\n$$text\nx = y\n$$",
+        [
+            P(2, E.MISSING_NEWLINE_AFTER_OPENING_DISPLAY),
+            P(4, E.MISSING_NEWLINE_AFTER_OPENING_DISPLAY),
+        ],
+    ),
+    (
+        "Expression:\n$$\nx = y\ntext$$",
+        [P(4, E.MISSING_NEWLINE_BEFORE_CLOSING_DISPLAY)],
+    ),
+    (
+        "Expression $$x = y$.",
+        [
+            P(1, E.MISSING_NEWLINE_BEFORE_OPENING_DISPLAY),
+            P(1, E.MISSING_CLOSING_SINGLE_DOLLAR),
+        ],
+    ),
+    (
+        "Expression $x = y$$",
+        [
+            P(1, E.DOUBLE_DOLLAR_INSTEAD_OF_CLOSING_SINGLE),
+            P(1, E.MISSING_CLOSING_SINGLE_DOLLAR),
+        ],
+    ),
+    (
+        "Expression $$x = y$",
+        [
+            P(1, E.MISSING_NEWLINE_BEFORE_OPENING_DISPLAY),
+            P(1, E.MISSING_CLOSING_SINGLE_DOLLAR),
+        ],
+    ),
+    # An unclosed "$" must report MISSING_CLOSING_SINGLE_DOLLAR, not a
+    # newline-related error, however many lines it spans before EOF.
+    ("This $x is unclosed.\n", [P(1, E.MISSING_CLOSING_SINGLE_DOLLAR)]),
+    # Two independent problems on different lines are both reported.
+    (
+        "Expression $x = y$$.\nAnother $a = b$$.\n",
+        [
+            P(1, E.DOUBLE_DOLLAR_INSTEAD_OF_CLOSING_SINGLE),
+            P(2, E.MISSING_NEWLINE_BEFORE_OPENING_DISPLAY),
+        ],
+    ),
+]
+
+
+@pytest.mark.parametrize("content", VALID)
+def test_valid_markdown_passes(content: str) -> None:
+    assert math_delimiter_checker(content) == []
+    assert check_markdown(content) == []
+
+
+@pytest.mark.parametrize("content, expected", INVALID)
+def test_invalid_markdown_is_reported(
+    content: str, expected: list[MathDelimiterProblem]
+) -> None:
+    assert math_delimiter_checker(content) == expected
+    assert check_markdown(content) == expected