diff --git a/shared/static/scripts/sx-test-runner.js b/shared/static/scripts/sx-test-runner.js
new file mode 100644
index 0000000..89ce30c
--- /dev/null
+++ b/shared/static/scripts/sx-test-runner.js
@@ -0,0 +1,96 @@
+// sx-test-runner.js — Run test.sx in the browser using sx-browser.js.
+// Loaded on the /specs/testing page. Uses the Sx global.
+(function() {
+  var NIL = Sx.NIL;
+  function isNil(x) { return x === NIL || x === null || x === undefined; }
+  function deepEqual(a, b) {
+    if (a === b) return true;
+    if (isNil(a) && isNil(b)) return true;
+    if (typeof a !== typeof b) return false;
+    if (Array.isArray(a) && Array.isArray(b)) {
+      if (a.length !== b.length) return false;
+      for (var i = 0; i < a.length; i++) if (!deepEqual(a[i], b[i])) return false;
+      return true;
+    }
+    if (a && typeof a === "object" && b && typeof b === "object") {
+      var ka = Object.keys(a), kb = Object.keys(b);
+      if (ka.length !== kb.length) return false;
+      for (var j = 0; j < ka.length; j++) if (!deepEqual(a[ka[j]], b[ka[j]])) return false;
+      return true;
+    }
+    return false;
+  }
+
+  window.sxRunTests = function(srcId, outId, btnId) {
+    var src = document.getElementById(srcId).textContent;
+    var out = document.getElementById(outId);
+    var btn = document.getElementById(btnId);
+
+    var stack = [], passed = 0, failed = 0, num = 0, lines = [];
+
+    var env = {
+      "try-call": function(thunk) {
+        try {
+          Sx.eval([thunk], env);
+          return { ok: true };
+        } catch(e) {
+          return { ok: false, error: e.message || String(e) };
+        }
+      },
+      "report-pass": function(name) {
+        num++; passed++;
+        lines.push("ok " + num + " - " + stack.concat([name]).join(" > "));
+      },
+      "report-fail": function(name, error) {
+        num++; failed++;
+        lines.push("not ok " + num + " - " + stack.concat([name]).join(" > "));
+        lines.push("  # " + error);
+      },
+      "push-suite": function(name) { stack.push(name); },
+      "pop-suite": function() { stack.pop(); },
+
+      "equal?": function(a, b) { return deepEqual(a, b); },
+      "eq?": function(a, b) { return a === b; },
+      "boolean?": function(x) { return typeof x === "boolean"; },
+      "string-length": function(s) { return String(s).length; },
+      "substring": function(s, start, end) { return String(s).slice(start, end); },
+      "string-contains?": function(s, n) { return String(s).indexOf(n) !== -1; },
+      "upcase": function(s) { return String(s).toUpperCase(); },
+      "downcase": function(s) { return String(s).toLowerCase(); },
+      "reverse": function(c) { return c ? c.slice().reverse() : []; },
+      "flatten": function(c) {
+        var r = [];
+        for (var i = 0; i < (c||[]).length; i++) {
+          if (Array.isArray(c[i])) for (var j = 0; j < c[i].length; j++) r.push(c[i][j]);
+          else r.push(c[i]);
+        }
+        return r;
+      },
+      "has-key?": function(d, k) { return d && typeof d === "object" && k in d; },
+      "append": function(c, x) { return Array.isArray(x) ? (c||[]).concat(x) : (c||[]).concat([x]); },
+    };
+
+    try {
+      var t0 = performance.now();
+      var exprs = Sx.parseAll(src);
+      for (var i = 0; i < exprs.length; i++) Sx.eval(exprs[i], env);
+      var elapsed = Math.round(performance.now() - t0);
+      lines.push("");
+      lines.push("1.." + num);
+      lines.push("# tests " + (passed + failed));
+      lines.push("# pass  " + passed);
+      if (failed > 0) lines.push("# fail  " + failed);
+      lines.push("# time  " + elapsed + "ms");
+    } catch(e) {
+      lines.push("");
+      lines.push("FATAL: " + (e.message || String(e)));
+    }
+
+    out.textContent = lines.join("\n");
+    out.style.display = "block";
+    btn.textContent = passed + "/" + (passed + failed) + " passed" + (failed === 0 ? "" : " (" + failed + " failed)");
+    btn.className = failed > 0
+      ? "px-4 py-2 rounded-md bg-red-600 text-white font-medium text-sm cursor-default"
+      : "px-4 py-2 rounded-md bg-green-600 text-white font-medium text-sm cursor-default";
+  };
+})();
diff --git a/shared/sx/primitives.py b/shared/sx/primitives.py
index cbd356d..f3ff5a1 100644
--- a/shared/sx/primitives.py
+++ b/shared/sx/primitives.py
@@ -192,9 +192,13 @@ def prim_is_zero(n: Any) -> bool:
 def prim_is_nil(x: Any) -> bool:
     return x is None or x is NIL
 
+@register_primitive("boolean?")
+def prim_is_boolean(x: Any) -> bool:
+    return isinstance(x, bool)
+
 @register_primitive("number?")
 def prim_is_number(x: Any) -> bool:
-    return isinstance(x, (int, float))
+    return isinstance(x, (int, float)) and not isinstance(x, bool)
 
 @register_primitive("string?")
 def prim_is_string(x: Any) -> bool:
@@ -268,13 +272,27 @@ def prim_concat(*colls: Any) -> list:
     return result
 
 @register_primitive("upper")
+@register_primitive("upcase")
 def prim_upper(s: str) -> str:
     return s.upper()
 
 @register_primitive("lower")
+@register_primitive("downcase")
 def prim_lower(s: str) -> str:
     return s.lower()
 
+@register_primitive("string-length")
+def prim_string_length(s: str) -> int:
+    return len(s)
+
+@register_primitive("substring")
+def prim_substring(s: str, start: int, end: int) -> str:
+    return s[int(start):int(end)]
+
+@register_primitive("string-contains?")
+def prim_string_contains(s: str, needle: str) -> bool:
+    return needle in s
+
 @register_primitive("trim")
 def prim_trim(s: str) -> str:
     return s.strip()
@@ -384,8 +402,31 @@ def prim_cons(x: Any, coll: Any) -> list:
 
 @register_primitive("append")
 def prim_append(coll: Any, x: Any) -> list:
+    if isinstance(x, list):
+        return list(coll) + x if coll else list(x)
     return list(coll) + [x] if coll else [x]
 
+@register_primitive("reverse")
+def prim_reverse(coll: Any) -> list:
+    return list(reversed(coll)) if coll else []
+
+@register_primitive("flatten")
+def prim_flatten(coll: Any) -> list:
+    result = []
+    for item in (coll or []):
+        if isinstance(item, list):
+            result.extend(item)
+        else:
+            result.append(item)
+    return result
+
+@register_primitive("has-key?")
+def prim_has_key(d: Any, key: Any) -> bool:
+    if not isinstance(d, dict):
+        return False
+    k = key.name if isinstance(key, Keyword) else key
+    return k in d
+
 @register_primitive("append!")
 def prim_append_mut(coll: Any, x: Any) -> list:
     coll.append(x)
diff --git a/shared/sx/ref/bootstrap_test.py b/shared/sx/ref/bootstrap_test.py
new file mode 100644
index 0000000..08eb64b
--- /dev/null
+++ b/shared/sx/ref/bootstrap_test.py
@@ -0,0 +1,245 @@
+#!/usr/bin/env python3
+"""
+Bootstrap compiler: test.sx -> pytest test module.
+
+Reads test.sx and emits a Python test file that runs each deftest
+as a pytest test case, grouped into classes by defsuite.
+
+The emitted tests use the SX evaluator to run SX test bodies,
+verifying that the Python implementation matches the spec.
+
+Usage:
+    python bootstrap_test.py --output shared/sx/tests/test_sx_spec.py
+    pytest shared/sx/tests/test_sx_spec.py -v
+"""
+from __future__ import annotations
+
+import os
+import re
+import sys
+import argparse
+
+_HERE = os.path.dirname(os.path.abspath(__file__))
+_PROJECT = os.path.abspath(os.path.join(_HERE, "..", "..", ".."))
+sys.path.insert(0, _PROJECT)
+
+from shared.sx.parser import parse_all
+from shared.sx.types import Symbol, Keyword, NIL as SX_NIL
+
+
+def _slugify(name: str) -> str:
+    """Convert a test/suite name to a valid Python identifier."""
+    s = name.lower().strip()
+    s = re.sub(r'[^a-z0-9]+', '_', s)
+    s = s.strip('_')
+    return s
+
+
+def _sx_to_source(expr) -> str:
+    """Convert an SX AST node back to SX source string."""
+    if isinstance(expr, bool):
+        return "true" if expr else "false"
+    if isinstance(expr, (int, float)):
+        return str(expr)
+    if isinstance(expr, str):
+        escaped = expr.replace('\\', '\\\\').replace('"', '\\"')
+        return f'"{escaped}"'
+    if expr is None or expr is SX_NIL:
+        return "nil"
+    if isinstance(expr, Symbol):
+        return expr.name
+    if isinstance(expr, Keyword):
+        return f":{expr.name}"
+    if isinstance(expr, dict):
+        pairs = []
+        for k, v in expr.items():
+            pairs.append(f":{k} {_sx_to_source(v)}")
+        return "{" + " ".join(pairs) + "}"
+    if isinstance(expr, list):
+        if not expr:
+            return "()"
+        return "(" + " ".join(_sx_to_source(e) for e in expr) + ")"
+    return str(expr)
+
+
+def _parse_test_sx(path: str) -> tuple[list[dict], list]:
+    """Parse test.sx and return (suites, preamble_exprs).
+
+    Preamble exprs are define forms (assertion helpers) that must be
+    evaluated before tests run. Suites contain the actual test cases.
+    """
+    with open(path) as f:
+        content = f.read()
+
+    exprs = parse_all(content)
+    suites = []
+    preamble = []
+
+    for expr in exprs:
+        if not isinstance(expr, list) or not expr:
+            continue
+        head = expr[0]
+        if isinstance(head, Symbol) and head.name == "defsuite":
+            suite = _parse_suite(expr)
+            if suite:
+                suites.append(suite)
+        elif isinstance(head, Symbol) and head.name == "define":
+            preamble.append(expr)
+
+    return suites, preamble
+
+
+def _parse_suite(expr: list) -> dict | None:
+    """Parse a (defsuite "name" ...) form."""
+    if len(expr) < 2:
+        return None
+
+    name = expr[1]
+    if not isinstance(name, str):
+        return None
+
+    tests = []
+    for child in expr[2:]:
+        if not isinstance(child, list) or not child:
+            continue
+        head = child[0]
+        if isinstance(head, Symbol):
+            if head.name == "deftest":
+                test = _parse_test(child)
+                if test:
+                    tests.append(test)
+            elif head.name == "defsuite":
+                sub = _parse_suite(child)
+                if sub:
+                    tests.append(sub)
+
+    return {"type": "suite", "name": name, "tests": tests}
+
+
+def _parse_test(expr: list) -> dict | None:
+    """Parse a (deftest "name" body ...) form."""
+    if len(expr) < 3:
+        return None
+    name = expr[1]
+    if not isinstance(name, str):
+        return None
+    body = expr[2:]
+    return {"type": "test", "name": name, "body": body}
+
+
+def _emit_py(suites: list[dict], preamble: list) -> str:
+    """Emit a pytest module from parsed suites."""
+    # Serialize preamble (assertion helpers) as SX source
+    preamble_sx = "\n".join(_sx_to_source(expr) for expr in preamble)
+    preamble_escaped = preamble_sx.replace('\\', '\\\\').replace("'", "\\'")
+
+    lines = []
+    lines.append('"""Auto-generated from test.sx — SX spec self-tests.')
+    lines.append('')
+    lines.append('DO NOT EDIT. Regenerate with:')
+    lines.append('    python shared/sx/ref/bootstrap_test.py --output shared/sx/tests/test_sx_spec.py')
+    lines.append('"""')
+    lines.append('from __future__ import annotations')
+    lines.append('')
+    lines.append('import pytest')
+    lines.append('from shared.sx.parser import parse_all')
+    lines.append('from shared.sx.evaluator import _eval, _trampoline')
+    lines.append('')
+    lines.append('')
+    lines.append(f"_PREAMBLE = '''{preamble_escaped}'''")
+    lines.append('')
+    lines.append('')
+    lines.append('def _make_env() -> dict:')
+    lines.append('    """Create a fresh env with assertion helpers loaded."""')
+    lines.append('    env = {}')
+    lines.append('    for expr in parse_all(_PREAMBLE):')
+    lines.append('        _trampoline(_eval(expr, env))')
+    lines.append('    return env')
+    lines.append('')
+    lines.append('')
+    lines.append('def _run(sx_source: str, env: dict | None = None) -> object:')
+    lines.append('    """Evaluate SX source and return the result."""')
+    lines.append('    if env is None:')
+    lines.append('        env = _make_env()')
+    lines.append('    exprs = parse_all(sx_source)')
+    lines.append('    result = None')
+    lines.append('    for expr in exprs:')
+    lines.append('        result = _trampoline(_eval(expr, env))')
+    lines.append('    return result')
+    lines.append('')
+
+    for suite in suites:
+        _emit_suite(suite, lines, indent=0)
+
+    return "\n".join(lines)
+
+
+def _emit_suite(suite: dict, lines: list[str], indent: int):
+    """Emit a pytest class for a suite."""
+    class_name = f"TestSpec{_slugify(suite['name']).title().replace('_', '')}"
+    pad = "    " * indent
+    lines.append(f'{pad}class {class_name}:')
+    lines.append(f'{pad}    """test.sx suite: {suite["name"]}"""')
+    lines.append('')
+
+    for item in suite["tests"]:
+        if item["type"] == "test":
+            _emit_test(item, lines, indent + 1)
+        elif item["type"] == "suite":
+            _emit_suite(item, lines, indent + 1)
+
+    lines.append('')
+
+
+def _emit_test(test: dict, lines: list[str], indent: int):
+    """Emit a pytest test method."""
+    method_name = f"test_{_slugify(test['name'])}"
+    pad = "    " * indent
+
+    # Convert body expressions to SX source
+    body_parts = []
+    for expr in test["body"]:
+        body_parts.append(_sx_to_source(expr))
+
+    # Wrap in (do ...) if multiple expressions, or use single
+    if len(body_parts) == 1:
+        sx_source = body_parts[0]
+    else:
+        sx_source = "(do " + " ".join(body_parts) + ")"
+
+    # Escape for Python string
+    sx_escaped = sx_source.replace('\\', '\\\\').replace("'", "\\'")
+
+    lines.append(f"{pad}def {method_name}(self):")
+    lines.append(f"{pad}    _run('{sx_escaped}')")
+    lines.append('')
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Bootstrap test.sx to pytest")
+    parser.add_argument("--output", "-o", help="Output file path")
+    parser.add_argument("--dry-run", action="store_true", help="Print to stdout")
+    args = parser.parse_args()
+
+    test_sx = os.path.join(_HERE, "test.sx")
+    suites, preamble = _parse_test_sx(test_sx)
+
+    print(f"Parsed {len(suites)} suites, {len(preamble)} preamble defines from test.sx", file=sys.stderr)
+    total_tests = sum(
+        sum(1 for t in s["tests"] if t["type"] == "test")
+        for s in suites
+    )
+    print(f"Total test cases: {total_tests}", file=sys.stderr)
+
+    output = _emit_py(suites, preamble)
+
+    if args.output and not args.dry_run:
+        with open(args.output, "w") as f:
+            f.write(output)
+        print(f"Wrote {args.output}", file=sys.stderr)
+    else:
+        print(output)
+
+
+if __name__ == "__main__":
+    main()
diff --git a/shared/sx/ref/primitives.sx b/shared/sx/ref/primitives.sx
index 6035a66..170037d 100644
--- a/shared/sx/ref/primitives.sx
+++ b/shared/sx/ref/primitives.sx
@@ -208,10 +208,16 @@
   :returns "boolean"
   :doc "True if x is nil/null/None.")
 
+(define-primitive "boolean?"
+  :params (x)
+  :returns "boolean"
+  :doc "True if x is a boolean (true or false). Must be checked before
+        number? on platforms where booleans are numeric subtypes.")
+
 (define-primitive "number?"
   :params (x)
   :returns "boolean"
-  :doc "True if x is a number (int or float).")
+  :doc "True if x is a number (int or float). Excludes booleans.")
 
 (define-primitive "string?"
   :params (x)
@@ -277,11 +283,36 @@
   :returns "string"
   :doc "Uppercase string.")
 
+(define-primitive "upcase"
+  :params (s)
+  :returns "string"
+  :doc "Alias for upper. Uppercase string.")
+
 (define-primitive "lower"
   :params (s)
   :returns "string"
   :doc "Lowercase string.")
 
+(define-primitive "downcase"
+  :params (s)
+  :returns "string"
+  :doc "Alias for lower. Lowercase string.")
+
+(define-primitive "string-length"
+  :params (s)
+  :returns "number"
+  :doc "Length of string in characters.")
+
+(define-primitive "substring"
+  :params (s start end)
+  :returns "string"
+  :doc "Extract substring from start (inclusive) to end (exclusive).")
+
+(define-primitive "string-contains?"
+  :params (s needle)
+  :returns "boolean"
+  :doc "True if string s contains substring needle.")
+
 (define-primitive "trim"
   :params (s)
   :returns "string"
@@ -382,13 +413,23 @@
 (define-primitive "append"
   :params (coll x)
   :returns "list"
-  :doc "Append x to end of coll (returns new list).")
+  :doc "If x is a list, concatenate. Otherwise append x as single element.")
 
 (define-primitive "append!"
   :params (coll x)
   :returns "list"
   :doc "Mutate coll by appending x in-place. Returns coll.")
 
+(define-primitive "reverse"
+  :params (coll)
+  :returns "list"
+  :doc "Return coll in reverse order.")
+
+(define-primitive "flatten"
+  :params (coll)
+  :returns "list"
+  :doc "Flatten one level of nesting. Nested lists become top-level elements.")
+
 (define-primitive "chunk-every"
   :params (coll n)
   :returns "list"
@@ -421,6 +462,11 @@
   :returns "dict"
   :doc "Merge dicts left to right. Later keys win. Skips nil.")
 
+(define-primitive "has-key?"
+  :params (d key)
+  :returns "boolean"
+  :doc "True if dict d contains key.")
+
 (define-primitive "assoc"
   :params (d &rest pairs)
   :returns "dict"
diff --git a/shared/sx/ref/test.sx b/shared/sx/ref/test.sx
new file mode 100644
index 0000000..2f6fd97
--- /dev/null
+++ b/shared/sx/ref/test.sx
@@ -0,0 +1,597 @@
+;; ==========================================================================
+;; test.sx — Self-hosting SX test framework
+;;
+;; Defines a minimal test framework in SX that tests SX — the language
+;; proves its own correctness. The framework is self-executing: any host
+;; that provides 5 platform functions can evaluate this file directly.
+;;
+;; Platform functions required:
+;;   try-call     (thunk)        → {:ok true} | {:ok false :error "msg"}
+;;   report-pass  (name)         → platform-specific pass output
+;;   report-fail  (name error)   → platform-specific fail output
+;;   push-suite   (name)         → push suite name onto context stack
+;;   pop-suite    ()             → pop suite name from context stack
+;;
+;; Usage:
+;;   ;; Host injects platform functions into env, then:
+;;   (eval-file "test.sx" env)
+;;
+;; The same test.sx runs on every host — Python, JavaScript, etc.
+;; ==========================================================================
+
+
+;; --------------------------------------------------------------------------
+;; 1. Test framework macros
+;; --------------------------------------------------------------------------
+;;
+;; deftest and defsuite are macros that make test.sx directly executable.
+;; The host provides try-call (error catching), reporting, and suite
+;; context — everything else is pure SX.
+
+(defmacro deftest (name &rest body)
+  `(let ((result (try-call (fn () ,@body))))
+     (if (get result "ok")
+       (report-pass ,name)
+       (report-fail ,name (get result "error")))))
+
+(defmacro defsuite (name &rest items)
+  `(do (push-suite ,name)
+       ,@items
+       (pop-suite)))
+
+
+;; --------------------------------------------------------------------------
+;; 2. Assertion helpers — defined in SX, available in test bodies
+;; --------------------------------------------------------------------------
+;;
+;; These are regular functions (not special forms). They use the `assert`
+;; primitive underneath but provide better error messages.
+
+(define assert-equal
+  (fn (expected actual)
+    (assert (equal? expected actual)
+      (str "Expected " (str expected) " but got " (str actual)))))
+
+(define assert-not-equal
+  (fn (a b)
+    (assert (not (equal? a b))
+      (str "Expected values to differ but both are " (str a)))))
+
+(define assert-true
+  (fn (val)
+    (assert val (str "Expected truthy but got " (str val)))))
+
+(define assert-false
+  (fn (val)
+    (assert (not val) (str "Expected falsy but got " (str val)))))
+
+(define assert-nil
+  (fn (val)
+    (assert (nil? val) (str "Expected nil but got " (str val)))))
+
+(define assert-type
+  (fn (expected-type val)
+    ;; Implemented via predicate dispatch since type-of is a platform
+    ;; function not available in all hosts. Uses nested if to avoid
+    ;; Scheme-style cond detection for 2-element predicate calls.
+    ;; Boolean checked before number (subtypes on some platforms).
+    (let ((actual-type
+            (if (nil? val) "nil"
+              (if (boolean? val) "boolean"
+                (if (number? val) "number"
+                  (if (string? val) "string"
+                    (if (list? val) "list"
+                      (if (dict? val) "dict"
+                        "unknown"))))))))
+      (assert (= expected-type actual-type)
+        (str "Expected type " expected-type " but got " actual-type)))))
+
+(define assert-length
+  (fn (expected-len col)
+    (assert (= (len col) expected-len)
+      (str "Expected length " expected-len " but got " (len col)))))
+
+(define assert-contains
+  (fn (item col)
+    (assert (some (fn (x) (equal? x item)) col)
+      (str "Expected collection to contain " (str item)))))
+
+(define assert-throws
+  (fn (thunk)
+    (let ((result (try-call thunk)))
+      (assert (not (get result "ok"))
+        "Expected an error to be thrown but none was"))))
+
+
+;; ==========================================================================
+;; 3. Test suites — SX testing SX
+;; ==========================================================================
+
+
+;; --------------------------------------------------------------------------
+;; 3a. Literals and types
+;; --------------------------------------------------------------------------
+
+(defsuite "literals"
+  (deftest "numbers are numbers"
+    (assert-type "number" 42)
+    (assert-type "number" 3.14)
+    (assert-type "number" -1))
+
+  (deftest "strings are strings"
+    (assert-type "string" "hello")
+    (assert-type "string" ""))
+
+  (deftest "booleans are booleans"
+    (assert-type "boolean" true)
+    (assert-type "boolean" false))
+
+  (deftest "nil is nil"
+    (assert-type "nil" nil)
+    (assert-nil nil))
+
+  (deftest "lists are lists"
+    (assert-type "list" (list 1 2 3))
+    (assert-type "list" (list)))
+
+  (deftest "dicts are dicts"
+    (assert-type "dict" {:a 1 :b 2})))
+
+
+;; --------------------------------------------------------------------------
+;; 3b. Arithmetic
+;; --------------------------------------------------------------------------
+
+(defsuite "arithmetic"
+  (deftest "addition"
+    (assert-equal 3 (+ 1 2))
+    (assert-equal 0 (+ 0 0))
+    (assert-equal -1 (+ 1 -2))
+    (assert-equal 10 (+ 1 2 3 4)))
+
+  (deftest "subtraction"
+    (assert-equal 1 (- 3 2))
+    (assert-equal -1 (- 2 3)))
+
+  (deftest "multiplication"
+    (assert-equal 6 (* 2 3))
+    (assert-equal 0 (* 0 100))
+    (assert-equal 24 (* 1 2 3 4)))
+
+  (deftest "division"
+    (assert-equal 2 (/ 6 3))
+    (assert-equal 2.5 (/ 5 2)))
+
+  (deftest "modulo"
+    (assert-equal 1 (mod 7 3))
+    (assert-equal 0 (mod 6 3))))
+
+
+;; --------------------------------------------------------------------------
+;; 3c. Comparison
+;; --------------------------------------------------------------------------
+
+(defsuite "comparison"
+  (deftest "equality"
+    (assert-true (= 1 1))
+    (assert-false (= 1 2))
+    (assert-true (= "a" "a"))
+    (assert-false (= "a" "b")))
+
+  (deftest "deep equality"
+    (assert-true (equal? (list 1 2 3) (list 1 2 3)))
+    (assert-false (equal? (list 1 2) (list 1 3)))
+    (assert-true (equal? {:a 1} {:a 1}))
+    (assert-false (equal? {:a 1} {:a 2})))
+
+  (deftest "ordering"
+    (assert-true (< 1 2))
+    (assert-false (< 2 1))
+    (assert-true (> 2 1))
+    (assert-true (<= 1 1))
+    (assert-true (<= 1 2))
+    (assert-true (>= 2 2))
+    (assert-true (>= 3 2))))
+
+
+;; --------------------------------------------------------------------------
+;; 3d. String operations
+;; --------------------------------------------------------------------------
+
+(defsuite "strings"
+  (deftest "str concatenation"
+    (assert-equal "abc" (str "a" "b" "c"))
+    (assert-equal "hello world" (str "hello" " " "world"))
+    (assert-equal "42" (str 42))
+    (assert-equal "" (str)))
+
+  (deftest "string-length"
+    (assert-equal 5 (string-length "hello"))
+    (assert-equal 0 (string-length "")))
+
+  (deftest "substring"
+    (assert-equal "ell" (substring "hello" 1 4))
+    (assert-equal "hello" (substring "hello" 0 5)))
+
+  (deftest "string-contains?"
+    (assert-true (string-contains? "hello world" "world"))
+    (assert-false (string-contains? "hello" "xyz")))
+
+  (deftest "upcase and downcase"
+    (assert-equal "HELLO" (upcase "hello"))
+    (assert-equal "hello" (downcase "HELLO")))
+
+  (deftest "trim"
+    (assert-equal "hello" (trim "  hello  "))
+    (assert-equal "hello" (trim "hello")))
+
+  (deftest "split and join"
+    (assert-equal (list "a" "b" "c") (split "a,b,c" ","))
+    (assert-equal "a-b-c" (join "-" (list "a" "b" "c")))))
+
+
+;; --------------------------------------------------------------------------
+;; 3e. List operations
+;; --------------------------------------------------------------------------
+
+(defsuite "lists"
+  (deftest "constructors"
+    (assert-equal (list 1 2 3) (list 1 2 3))
+    (assert-equal (list) (list))
+    (assert-length 3 (list 1 2 3)))
+
+  (deftest "first and rest"
+    (assert-equal 1 (first (list 1 2 3)))
+    (assert-equal (list 2 3) (rest (list 1 2 3)))
+    (assert-nil (first (list)))
+    (assert-equal (list) (rest (list))))
+
+  (deftest "nth"
+    (assert-equal 1 (nth (list 1 2 3) 0))
+    (assert-equal 2 (nth (list 1 2 3) 1))
+    (assert-equal 3 (nth (list 1 2 3) 2)))
+
+  (deftest "last"
+    (assert-equal 3 (last (list 1 2 3)))
+    (assert-nil (last (list))))
+
+  (deftest "cons and append"
+    (assert-equal (list 0 1 2) (cons 0 (list 1 2)))
+    (assert-equal (list 1 2 3 4) (append (list 1 2) (list 3 4))))
+
+  (deftest "reverse"
+    (assert-equal (list 3 2 1) (reverse (list 1 2 3)))
+    (assert-equal (list) (reverse (list))))
+
+  (deftest "empty?"
+    (assert-true (empty? (list)))
+    (assert-false (empty? (list 1))))
+
+  (deftest "len"
+    (assert-equal 0 (len (list)))
+    (assert-equal 3 (len (list 1 2 3))))
+
+  (deftest "contains?"
+    (assert-true (contains? (list 1 2 3) 2))
+    (assert-false (contains? (list 1 2 3) 4)))
+
+  (deftest "flatten"
+    (assert-equal (list 1 2 3 4) (flatten (list (list 1 2) (list 3 4))))))
+
+
+;; --------------------------------------------------------------------------
+;; 3f. Dict operations
+;; --------------------------------------------------------------------------
+
+(defsuite "dicts"
+  (deftest "dict literal"
+    (assert-type "dict" {:a 1 :b 2})
+    (assert-equal 1 (get {:a 1} "a"))
+    (assert-equal 2 (get {:a 1 :b 2} "b")))
+
+  (deftest "assoc"
+    (assert-equal {:a 1 :b 2} (assoc {:a 1} "b" 2))
+    (assert-equal {:a 99} (assoc {:a 1} "a" 99)))
+
+  (deftest "dissoc"
+    (assert-equal {:b 2} (dissoc {:a 1 :b 2} "a")))
+
+  (deftest "keys and vals"
+    (let ((d {:a 1 :b 2}))
+      (assert-length 2 (keys d))
+      (assert-length 2 (vals d))
+      (assert-contains "a" (keys d))
+      (assert-contains "b" (keys d))))
+
+  (deftest "has-key?"
+    (assert-true (has-key? {:a 1} "a"))
+    (assert-false (has-key? {:a 1} "b")))
+
+  (deftest "merge"
+    (assert-equal {:a 1 :b 2 :c 3}
+      (merge {:a 1 :b 2} {:c 3}))
+    (assert-equal {:a 99 :b 2}
+      (merge {:a 1 :b 2} {:a 99}))))
+
+
+;; --------------------------------------------------------------------------
+;; 3g. Predicates
+;; --------------------------------------------------------------------------
+
+(defsuite "predicates"
+  (deftest "nil?"
+    (assert-true (nil? nil))
+    (assert-false (nil? 0))
+    (assert-false (nil? false))
+    (assert-false (nil? "")))
+
+  (deftest "number?"
+    (assert-true (number? 42))
+    (assert-true (number? 3.14))
+    (assert-false (number? "42")))
+
+  (deftest "string?"
+    (assert-true (string? "hello"))
+    (assert-false (string? 42)))
+
+  (deftest "list?"
+    (assert-true (list? (list 1 2)))
+    (assert-false (list? "not a list")))
+
+  (deftest "dict?"
+    (assert-true (dict? {:a 1}))
+    (assert-false (dict? (list 1))))
+
+  (deftest "boolean?"
+    (assert-true (boolean? true))
+    (assert-true (boolean? false))
+    (assert-false (boolean? nil))
+    (assert-false (boolean? 0)))
+
+  (deftest "not"
+    (assert-true (not false))
+    (assert-true (not nil))
+    (assert-false (not true))
+    (assert-false (not 1))
+    (assert-false (not "x"))))
+
+
+;; --------------------------------------------------------------------------
+;; 3h. Special forms
+;; --------------------------------------------------------------------------
+
+(defsuite "special-forms"
+  (deftest "if"
+    (assert-equal "yes" (if true "yes" "no"))
+    (assert-equal "no" (if false "yes" "no"))
+    (assert-equal "no" (if nil "yes" "no"))
+    (assert-nil (if false "yes")))
+
+  (deftest "when"
+    (assert-equal "yes" (when true "yes"))
+    (assert-nil (when false "yes")))
+
+  (deftest "cond"
+    (assert-equal "a" (cond true "a" :else "b"))
+    (assert-equal "b" (cond false "a" :else "b"))
+    (assert-equal "c" (cond
+                        false "a"
+                        false "b"
+                        :else "c")))
+
+  (deftest "and"
+    (assert-true (and true true))
+    (assert-false (and true false))
+    (assert-false (and false true))
+    (assert-equal 3 (and 1 2 3)))
+
+  (deftest "or"
+    (assert-equal 1 (or 1 2))
+    (assert-equal 2 (or false 2))
+    (assert-equal "fallback" (or nil false "fallback"))
+    (assert-false (or false false)))
+
+  (deftest "let"
+    (assert-equal 3 (let ((x 1) (y 2)) (+ x y)))
+    (assert-equal "hello world"
+      (let ((a "hello") (b " world")) (str a b))))
+
+  (deftest "let clojure-style"
+    (assert-equal 3 (let (x 1 y 2) (+ x y))))
+
+  (deftest "do / begin"
+    (assert-equal 3 (do 1 2 3))
+    (assert-equal "last" (begin "first" "middle" "last")))
+
+  (deftest "define"
+    (define x 42)
+    (assert-equal 42 x))
+
+  (deftest "set!"
+    (define x 1)
+    (set! x 2)
+    (assert-equal 2 x)))
+
+
+;; --------------------------------------------------------------------------
+;; 3i. Lambda and closures
+;; --------------------------------------------------------------------------
+
+(defsuite "lambdas"
+  (deftest "basic lambda"
+    (let ((add (fn (a b) (+ a b))))
+      (assert-equal 3 (add 1 2))))
+
+  (deftest "closure captures env"
+    (let ((x 10))
+      (let ((add-x (fn (y) (+ x y))))
+        (assert-equal 15 (add-x 5)))))
+
+  (deftest "lambda as argument"
+    (assert-equal (list 2 4 6)
+      (map (fn (x) (* x 2)) (list 1 2 3))))
+
+  (deftest "recursive lambda via define"
+    (define factorial
+      (fn (n) (if (<= n 1) 1 (* n (factorial (- n 1))))))
+    (assert-equal 120 (factorial 5)))
+
+  (deftest "higher-order returns lambda"
+    (let ((make-adder (fn (n) (fn (x) (+ n x)))))
+      (let ((add5 (make-adder 5)))
+        (assert-equal 8 (add5 3))))))
+
+
+;; --------------------------------------------------------------------------
+;; 3j. Higher-order forms
+;; --------------------------------------------------------------------------
+
+(defsuite "higher-order"
+  (deftest "map"
+    (assert-equal (list 2 4 6)
+      (map (fn (x) (* x 2)) (list 1 2 3)))
+    (assert-equal (list) (map (fn (x) x) (list))))
+
+  (deftest "filter"
+    (assert-equal (list 2 4)
+      (filter (fn (x) (= (mod x 2) 0)) (list 1 2 3 4)))
+    (assert-equal (list)
+      (filter (fn (x) false) (list 1 2 3))))
+
+  (deftest "reduce"
+    (assert-equal 10 (reduce (fn (acc x) (+ acc x)) 0 (list 1 2 3 4)))
+    (assert-equal 0 (reduce (fn (acc x) (+ acc x)) 0 (list))))
+
+  (deftest "some"
+    (assert-true (some (fn (x) (> x 3)) (list 1 2 3 4 5)))
+    (assert-false (some (fn (x) (> x 10)) (list 1 2 3))))
+
+  (deftest "every?"
+    (assert-true (every? (fn (x) (> x 0)) (list 1 2 3)))
+    (assert-false (every? (fn (x) (> x 2)) (list 1 2 3))))
+
+  (deftest "map-indexed"
+    (assert-equal (list "0:a" "1:b" "2:c")
+      (map-indexed (fn (i x) (str i ":" x)) (list "a" "b" "c")))))
+
+
+;; --------------------------------------------------------------------------
+;; 3k. Components
+;; --------------------------------------------------------------------------
+
+(defsuite "components"
+  (deftest "defcomp creates component"
+    (defcomp ~test-comp (&key title)
+      (div title))
+    ;; Component is bound and not nil
+    (assert-true (not (nil? ~test-comp))))
+
+  (deftest "component renders with keyword args"
+    (defcomp ~greeting (&key name)
+      (span (str "Hello, " name "!")))
+    (assert-true (not (nil? ~greeting))))
+
+  (deftest "component with children"
+    (defcomp ~box (&key &rest children)
+      (div :class "box" children))
+    (assert-true (not (nil? ~box))))
+
+  (deftest "component with default via or"
+    (defcomp ~label (&key text)
+      (span (or text "default")))
+    (assert-true (not (nil? ~label)))))
+
+
+;; --------------------------------------------------------------------------
+;; 3l. Macros
+;; --------------------------------------------------------------------------
+
+(defsuite "macros"
+  (deftest "defmacro creates macro"
+    (defmacro unless (cond &rest body)
+      `(if (not ,cond) (do ,@body)))
+    (assert-equal "yes" (unless false "yes"))
+    (assert-nil (unless true "no")))
+
+  (deftest "quasiquote and unquote"
+    (let ((x 42))
+      (assert-equal (list 1 42 3) `(1 ,x 3))))
+
+  (deftest "splice-unquote"
+    (let ((xs (list 2 3 4)))
+      (assert-equal (list 1 2 3 4 5) `(1 ,@xs 5)))))
+
+
+;; --------------------------------------------------------------------------
+;; 3m. Threading macro
+;; --------------------------------------------------------------------------
+
+(defsuite "threading"
+  (deftest "thread-first"
+    (assert-equal 8 (-> 5 (+ 1) (+ 2)))
+    (assert-equal "HELLO" (-> "hello" upcase))
+    (assert-equal "HELLO WORLD"
+      (-> "hello"
+          (str " world")
+          upcase))))
+
+
+;; --------------------------------------------------------------------------
+;; 3n. Truthiness
+;; --------------------------------------------------------------------------
+
+(defsuite "truthiness"
+  (deftest "truthy values"
+    (assert-true (if 1 true false))
+    (assert-true (if "x" true false))
+    (assert-true (if (list 1) true false))
+    (assert-true (if true true false)))
+
+  (deftest "falsy values"
+    (assert-false (if false true false))
+    (assert-false (if nil true false)))
+
+  ;; NOTE: empty list, zero, and empty string truthiness is
+  ;; platform-dependent. Python treats all three as falsy.
+  ;; JavaScript treats [] as truthy but 0 and "" as falsy.
+  ;; These tests are omitted — each bootstrapper should emit
+  ;; platform-specific truthiness tests instead.
+  )
+
+
+;; --------------------------------------------------------------------------
+;; 3o. Edge cases and regression tests
+;; --------------------------------------------------------------------------
+
+(defsuite "edge-cases"
+  (deftest "nested let scoping"
+    (let ((x 1))
+      (let ((x 2))
+        (assert-equal 2 x))
+      ;; outer x should be unchanged by inner let
+      ;; (this tests that let creates a new scope)
+      ))
+
+  (deftest "recursive map"
+    (assert-equal (list (list 2 4) (list 6 8))
+      (map (fn (sub) (map (fn (x) (* x 2)) sub))
+           (list (list 1 2) (list 3 4)))))
+
+  (deftest "keyword as value"
+    (assert-equal "class" :class)
+    (assert-equal "id" :id))
+
+  (deftest "dict with evaluated values"
+    (let ((x 42))
+      (assert-equal 42 (get {:val x} "val"))))
+
+  (deftest "nil propagation"
+    (assert-nil (get {:a 1} "missing"))
+    (assert-equal "default" (or (get {:a 1} "missing") "default")))
+
+  (deftest "empty operations"
+    (assert-equal (list) (map (fn (x) x) (list)))
+    (assert-equal (list) (filter (fn (x) true) (list)))
+    (assert-equal 0 (reduce (fn (acc x) (+ acc x)) 0 (list)))
+    (assert-equal 0 (len (list)))
+    (assert-equal "" (str))))
diff --git a/shared/sx/tests/run.js b/shared/sx/tests/run.js
new file mode 100644
index 0000000..ec77792
--- /dev/null
+++ b/shared/sx/tests/run.js
@@ -0,0 +1,108 @@
+// Run test.sx directly against sx-browser.js.
+//
+// sx-browser.js parses and evaluates test.sx — SX tests itself.
+// This script provides only platform functions (error catching, reporting).
+//
+// Usage: node shared/sx/tests/run.js
+
+Object.defineProperty(globalThis, "document", { value: undefined, writable: true });
+var path = require("path");
+var fs = require("fs");
+var Sx = require(path.resolve(__dirname, "../../static/scripts/sx-browser.js"));
+
+// --- Test state ---
+var suiteStack = [];
+var passed = 0, failed = 0, testNum = 0;
+
+// --- Helpers ---
+function isNil(x) { return x === Sx.NIL || x === null || x === undefined; }
+
+function deepEqual(a, b) {
+  if (a === b) return true;
+  if (isNil(a) && isNil(b)) return true;
+  if (typeof a !== typeof b) return false;
+  if (Array.isArray(a) && Array.isArray(b)) {
+    if (a.length !== b.length) return false;
+    for (var i = 0; i < a.length; i++) if (!deepEqual(a[i], b[i])) return false;
+    return true;
+  }
+  if (a && typeof a === "object" && b && typeof b === "object") {
+    var ka = Object.keys(a), kb = Object.keys(b);
+    if (ka.length !== kb.length) return false;
+    for (var j = 0; j < ka.length; j++) if (!deepEqual(a[ka[j]], b[ka[j]])) return false;
+    return true;
+  }
+  return false;
+}
+
+// --- Platform functions injected into the SX env ---
+var env = {
+  // Error catching — calls an SX thunk, returns result dict
+  "try-call": function(thunk) {
+    try {
+      Sx.eval([thunk], env);
+      return { ok: true };
+    } catch(e) {
+      return { ok: false, error: e.message || String(e) };
+    }
+  },
+
+  // Test reporting
+  "report-pass": function(name) {
+    testNum++;
+    passed++;
+    var fullName = suiteStack.concat([name]).join(" > ");
+    console.log("ok " + testNum + " - " + fullName);
+  },
+  "report-fail": function(name, error) {
+    testNum++;
+    failed++;
+    var fullName = suiteStack.concat([name]).join(" > ");
+    console.log("not ok " + testNum + " - " + fullName);
+    console.log("  # " + error);
+  },
+
+  // Suite context
+  "push-suite": function(name) { suiteStack.push(name); },
+  "pop-suite": function() { suiteStack.pop(); },
+
+  // Primitives that sx-browser.js has internally but doesn't expose through env
+  "equal?": function(a, b) { return deepEqual(a, b); },
+  "eq?": function(a, b) { return a === b; },
+  "boolean?": function(x) { return typeof x === "boolean"; },
+  "string-length": function(s) { return String(s).length; },
+  "substring": function(s, start, end) { return String(s).slice(start, end); },
+  "string-contains?": function(s, needle) { return String(s).indexOf(needle) !== -1; },
+  "upcase": function(s) { return String(s).toUpperCase(); },
+  "downcase": function(s) { return String(s).toLowerCase(); },
+  "reverse": function(c) { return c ? c.slice().reverse() : []; },
+  "flatten": function(c) {
+    var r = [];
+    for (var i = 0; i < (c||[]).length; i++) {
+      if (Array.isArray(c[i])) for (var j = 0; j < c[i].length; j++) r.push(c[i][j]);
+      else r.push(c[i]);
+    }
+    return r;
+  },
+  "has-key?": function(d, k) { return d && typeof d === "object" && k in d; },
+  "append": function(c, x) { return Array.isArray(x) ? (c||[]).concat(x) : (c||[]).concat([x]); },
+};
+
+// --- Read and evaluate test.sx ---
+var src = fs.readFileSync(path.resolve(__dirname, "../ref/test.sx"), "utf8");
+var exprs = Sx.parseAll(src);
+
+console.log("TAP version 13");
+for (var i = 0; i < exprs.length; i++) {
+  Sx.eval(exprs[i], env);
+}
+
+// --- Summary ---
+console.log("");
+console.log("1.." + testNum);
+console.log("# tests " + (passed + failed));
+console.log("# pass  " + passed);
+if (failed > 0) {
+  console.log("# fail  " + failed);
+  process.exit(1);
+}
diff --git a/shared/sx/tests/run.py b/shared/sx/tests/run.py
new file mode 100644
index 0000000..3fda836
--- /dev/null
+++ b/shared/sx/tests/run.py
@@ -0,0 +1,92 @@
+#!/usr/bin/env python3
+"""Run test.sx directly against the Python SX evaluator.
+
+The Python evaluator parses and evaluates test.sx — SX tests itself.
+This script provides only platform functions (error catching, reporting).
+
+Usage: python shared/sx/tests/run.py
+"""
+from __future__ import annotations
+
+import os
+import sys
+import traceback
+
+_HERE = os.path.dirname(os.path.abspath(__file__))
+_PROJECT = os.path.abspath(os.path.join(_HERE, "..", "..", ".."))
+sys.path.insert(0, _PROJECT)
+
+from shared.sx.parser import parse_all
+from shared.sx.evaluator import _eval, _trampoline
+
+# --- Test state ---
+suite_stack: list[str] = []
+passed = 0
+failed = 0
+test_num = 0
+
+
+def try_call(thunk):
+    """Call an SX thunk, catching errors."""
+    try:
+        _trampoline(_eval([thunk], {}))
+        return {"ok": True}
+    except Exception as e:
+        return {"ok": False, "error": str(e)}
+
+
+def report_pass(name):
+    global passed, test_num
+    test_num += 1
+    passed += 1
+    full_name = " > ".join(suite_stack + [name])
+    print(f"ok {test_num} - {full_name}")
+
+
+def report_fail(name, error):
+    global failed, test_num
+    test_num += 1
+    failed += 1
+    full_name = " > ".join(suite_stack + [name])
+    print(f"not ok {test_num} - {full_name}")
+    print(f"  # {error}")
+
+
+def push_suite(name):
+    suite_stack.append(name)
+
+
+def pop_suite():
+    suite_stack.pop()
+
+
+def main():
+    env = {
+        "try-call": try_call,
+        "report-pass": report_pass,
+        "report-fail": report_fail,
+        "push-suite": push_suite,
+        "pop-suite": pop_suite,
+    }
+
+    test_sx = os.path.join(_HERE, "..", "ref", "test.sx")
+    with open(test_sx) as f:
+        src = f.read()
+
+    exprs = parse_all(src)
+
+    print("TAP version 13")
+    for expr in exprs:
+        _trampoline(_eval(expr, env))
+
+    print()
+    print(f"1..{test_num}")
+    print(f"# tests {passed + failed}")
+    print(f"# pass  {passed}")
+    if failed > 0:
+        print(f"# fail  {failed}")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()
diff --git a/shared/sx/tests/test_sx_spec.py b/shared/sx/tests/test_sx_spec.py
new file mode 100644
index 0000000..db885cc
--- /dev/null
+++ b/shared/sx/tests/test_sx_spec.py
@@ -0,0 +1,343 @@
+"""Auto-generated from test.sx — SX spec self-tests.
+
+DO NOT EDIT. Regenerate with:
+    python shared/sx/ref/bootstrap_test.py --output shared/sx/tests/test_sx_spec.py
+"""
+from __future__ import annotations
+
+import pytest
+from shared.sx.parser import parse_all
+from shared.sx.evaluator import _eval, _trampoline
+
+
+_PREAMBLE = '''(define assert-equal (fn (expected actual) (assert (equal? expected actual) (str "Expected " (str expected) " but got " (str actual)))))
+(define assert-not-equal (fn (a b) (assert (not (equal? a b)) (str "Expected values to differ but both are " (str a)))))
+(define assert-true (fn (val) (assert val (str "Expected truthy but got " (str val)))))
+(define assert-false (fn (val) (assert (not val) (str "Expected falsy but got " (str val)))))
+(define assert-nil (fn (val) (assert (nil? val) (str "Expected nil but got " (str val)))))
+(define assert-type (fn (expected-type val) (let ((actual-type (if (nil? val) "nil" (if (boolean? val) "boolean" (if (number? val) "number" (if (string? val) "string" (if (list? val) "list" (if (dict? val) "dict" "unknown")))))))) (assert (= expected-type actual-type) (str "Expected type " expected-type " but got " actual-type)))))
+(define assert-length (fn (expected-len col) (assert (= (len col) expected-len) (str "Expected length " expected-len " but got " (len col)))))
+(define assert-contains (fn (item col) (assert (some (fn (x) (equal? x item)) col) (str "Expected collection to contain " (str item)))))
+(define assert-throws (fn (thunk) (let ((result (try-call thunk))) (assert (not (get result "ok")) "Expected an error to be thrown but none was"))))'''
+
+
+def _make_env() -> dict:
+    """Create a fresh env with assertion helpers loaded."""
+    env = {}
+    for expr in parse_all(_PREAMBLE):
+        _trampoline(_eval(expr, env))
+    return env
+
+
+def _run(sx_source: str, env: dict | None = None) -> object:
+    """Evaluate SX source and return the result."""
+    if env is None:
+        env = _make_env()
+    exprs = parse_all(sx_source)
+    result = None
+    for expr in exprs:
+        result = _trampoline(_eval(expr, env))
+    return result
+
+class TestSpecLiterals:
+    """test.sx suite: literals"""
+
+    def test_numbers_are_numbers(self):
+        _run('(do (assert-type "number" 42) (assert-type "number" 3.14) (assert-type "number" -1))')
+
+    def test_strings_are_strings(self):
+        _run('(do (assert-type "string" "hello") (assert-type "string" ""))')
+
+    def test_booleans_are_booleans(self):
+        _run('(do (assert-type "boolean" true) (assert-type "boolean" false))')
+
+    def test_nil_is_nil(self):
+        _run('(do (assert-type "nil" nil) (assert-nil nil))')
+
+    def test_lists_are_lists(self):
+        _run('(do (assert-type "list" (list 1 2 3)) (assert-type "list" (list)))')
+
+    def test_dicts_are_dicts(self):
+        _run('(assert-type "dict" {:a 1 :b 2})')
+
+
+class TestSpecArithmetic:
+    """test.sx suite: arithmetic"""
+
+    def test_addition(self):
+        _run('(do (assert-equal 3 (+ 1 2)) (assert-equal 0 (+ 0 0)) (assert-equal -1 (+ 1 -2)) (assert-equal 10 (+ 1 2 3 4)))')
+
+    def test_subtraction(self):
+        _run('(do (assert-equal 1 (- 3 2)) (assert-equal -1 (- 2 3)))')
+
+    def test_multiplication(self):
+        _run('(do (assert-equal 6 (* 2 3)) (assert-equal 0 (* 0 100)) (assert-equal 24 (* 1 2 3 4)))')
+
+    def test_division(self):
+        _run('(do (assert-equal 2 (/ 6 3)) (assert-equal 2.5 (/ 5 2)))')
+
+    def test_modulo(self):
+        _run('(do (assert-equal 1 (mod 7 3)) (assert-equal 0 (mod 6 3)))')
+
+
+class TestSpecComparison:
+    """test.sx suite: comparison"""
+
+    def test_equality(self):
+        _run('(do (assert-true (= 1 1)) (assert-false (= 1 2)) (assert-true (= "a" "a")) (assert-false (= "a" "b")))')
+
+    def test_deep_equality(self):
+        _run('(do (assert-true (equal? (list 1 2 3) (list 1 2 3))) (assert-false (equal? (list 1 2) (list 1 3))) (assert-true (equal? {:a 1} {:a 1})) (assert-false (equal? {:a 1} {:a 2})))')
+
+    def test_ordering(self):
+        _run('(do (assert-true (< 1 2)) (assert-false (< 2 1)) (assert-true (> 2 1)) (assert-true (<= 1 1)) (assert-true (<= 1 2)) (assert-true (>= 2 2)) (assert-true (>= 3 2)))')
+
+
+class TestSpecStrings:
+    """test.sx suite: strings"""
+
+    def test_str_concatenation(self):
+        _run('(do (assert-equal "abc" (str "a" "b" "c")) (assert-equal "hello world" (str "hello" " " "world")) (assert-equal "42" (str 42)) (assert-equal "" (str)))')
+
+    def test_string_length(self):
+        _run('(do (assert-equal 5 (string-length "hello")) (assert-equal 0 (string-length "")))')
+
+    def test_substring(self):
+        _run('(do (assert-equal "ell" (substring "hello" 1 4)) (assert-equal "hello" (substring "hello" 0 5)))')
+
+    def test_string_contains(self):
+        _run('(do (assert-true (string-contains? "hello world" "world")) (assert-false (string-contains? "hello" "xyz")))')
+
+    def test_upcase_and_downcase(self):
+        _run('(do (assert-equal "HELLO" (upcase "hello")) (assert-equal "hello" (downcase "HELLO")))')
+
+    def test_trim(self):
+        _run('(do (assert-equal "hello" (trim "  hello  ")) (assert-equal "hello" (trim "hello")))')
+
+    def test_split_and_join(self):
+        _run('(do (assert-equal (list "a" "b" "c") (split "a,b,c" ",")) (assert-equal "a-b-c" (join "-" (list "a" "b" "c"))))')
+
+
+class TestSpecLists:
+    """test.sx suite: lists"""
+
+    def test_constructors(self):
+        _run('(do (assert-equal (list 1 2 3) (list 1 2 3)) (assert-equal (list) (list)) (assert-length 3 (list 1 2 3)))')
+
+    def test_first_and_rest(self):
+        _run('(do (assert-equal 1 (first (list 1 2 3))) (assert-equal (list 2 3) (rest (list 1 2 3))) (assert-nil (first (list))) (assert-equal (list) (rest (list))))')
+
+    def test_nth(self):
+        _run('(do (assert-equal 1 (nth (list 1 2 3) 0)) (assert-equal 2 (nth (list 1 2 3) 1)) (assert-equal 3 (nth (list 1 2 3) 2)))')
+
+    def test_last(self):
+        _run('(do (assert-equal 3 (last (list 1 2 3))) (assert-nil (last (list))))')
+
+    def test_cons_and_append(self):
+        _run('(do (assert-equal (list 0 1 2) (cons 0 (list 1 2))) (assert-equal (list 1 2 3 4) (append (list 1 2) (list 3 4))))')
+
+    def test_reverse(self):
+        _run('(do (assert-equal (list 3 2 1) (reverse (list 1 2 3))) (assert-equal (list) (reverse (list))))')
+
+    def test_empty(self):
+        _run('(do (assert-true (empty? (list))) (assert-false (empty? (list 1))))')
+
+    def test_len(self):
+        _run('(do (assert-equal 0 (len (list))) (assert-equal 3 (len (list 1 2 3))))')
+
+    def test_contains(self):
+        _run('(do (assert-true (contains? (list 1 2 3) 2)) (assert-false (contains? (list 1 2 3) 4)))')
+
+    def test_flatten(self):
+        _run('(assert-equal (list 1 2 3 4) (flatten (list (list 1 2) (list 3 4))))')
+
+
+class TestSpecDicts:
+    """test.sx suite: dicts"""
+
+    def test_dict_literal(self):
+        _run('(do (assert-type "dict" {:a 1 :b 2}) (assert-equal 1 (get {:a 1} "a")) (assert-equal 2 (get {:a 1 :b 2} "b")))')
+
+    def test_assoc(self):
+        _run('(do (assert-equal {:a 1 :b 2} (assoc {:a 1} "b" 2)) (assert-equal {:a 99} (assoc {:a 1} "a" 99)))')
+
+    def test_dissoc(self):
+        _run('(assert-equal {:b 2} (dissoc {:a 1 :b 2} "a"))')
+
+    def test_keys_and_vals(self):
+        _run('(let ((d {:a 1 :b 2})) (assert-length 2 (keys d)) (assert-length 2 (vals d)) (assert-contains "a" (keys d)) (assert-contains "b" (keys d)))')
+
+    def test_has_key(self):
+        _run('(do (assert-true (has-key? {:a 1} "a")) (assert-false (has-key? {:a 1} "b")))')
+
+    def test_merge(self):
+        _run('(do (assert-equal {:a 1 :b 2 :c 3} (merge {:a 1 :b 2} {:c 3})) (assert-equal {:a 99 :b 2} (merge {:a 1 :b 2} {:a 99})))')
+
+
+class TestSpecPredicates:
+    """test.sx suite: predicates"""
+
+    def test_nil(self):
+        _run('(do (assert-true (nil? nil)) (assert-false (nil? 0)) (assert-false (nil? false)) (assert-false (nil? "")))')
+
+    def test_number(self):
+        _run('(do (assert-true (number? 42)) (assert-true (number? 3.14)) (assert-false (number? "42")))')
+
+    def test_string(self):
+        _run('(do (assert-true (string? "hello")) (assert-false (string? 42)))')
+
+    def test_list(self):
+        _run('(do (assert-true (list? (list 1 2))) (assert-false (list? "not a list")))')
+
+    def test_dict(self):
+        _run('(do (assert-true (dict? {:a 1})) (assert-false (dict? (list 1))))')
+
+    def test_boolean(self):
+        _run('(do (assert-true (boolean? true)) (assert-true (boolean? false)) (assert-false (boolean? nil)) (assert-false (boolean? 0)))')
+
+    def test_not(self):
+        _run('(do (assert-true (not false)) (assert-true (not nil)) (assert-false (not true)) (assert-false (not 1)) (assert-false (not "x")))')
+
+
+class TestSpecSpecialForms:
+    """test.sx suite: special-forms"""
+
+    def test_if(self):
+        _run('(do (assert-equal "yes" (if true "yes" "no")) (assert-equal "no" (if false "yes" "no")) (assert-equal "no" (if nil "yes" "no")) (assert-nil (if false "yes")))')
+
+    def test_when(self):
+        _run('(do (assert-equal "yes" (when true "yes")) (assert-nil (when false "yes")))')
+
+    def test_cond(self):
+        _run('(do (assert-equal "a" (cond true "a" :else "b")) (assert-equal "b" (cond false "a" :else "b")) (assert-equal "c" (cond false "a" false "b" :else "c")))')
+
+    def test_and(self):
+        _run('(do (assert-true (and true true)) (assert-false (and true false)) (assert-false (and false true)) (assert-equal 3 (and 1 2 3)))')
+
+    def test_or(self):
+        _run('(do (assert-equal 1 (or 1 2)) (assert-equal 2 (or false 2)) (assert-equal "fallback" (or nil false "fallback")) (assert-false (or false false)))')
+
+    def test_let(self):
+        _run('(do (assert-equal 3 (let ((x 1) (y 2)) (+ x y))) (assert-equal "hello world" (let ((a "hello") (b " world")) (str a b))))')
+
+    def test_let_clojure_style(self):
+        _run('(assert-equal 3 (let (x 1 y 2) (+ x y)))')
+
+    def test_do_begin(self):
+        _run('(do (assert-equal 3 (do 1 2 3)) (assert-equal "last" (begin "first" "middle" "last")))')
+
+    def test_define(self):
+        _run('(do (define x 42) (assert-equal 42 x))')
+
+    def test_set(self):
+        _run('(do (define x 1) (set! x 2) (assert-equal 2 x))')
+
+
+class TestSpecLambdas:
+    """test.sx suite: lambdas"""
+
+    def test_basic_lambda(self):
+        _run('(let ((add (fn (a b) (+ a b)))) (assert-equal 3 (add 1 2)))')
+
+    def test_closure_captures_env(self):
+        _run('(let ((x 10)) (let ((add-x (fn (y) (+ x y)))) (assert-equal 15 (add-x 5))))')
+
+    def test_lambda_as_argument(self):
+        _run('(assert-equal (list 2 4 6) (map (fn (x) (* x 2)) (list 1 2 3)))')
+
+    def test_recursive_lambda_via_define(self):
+        _run('(do (define factorial (fn (n) (if (<= n 1) 1 (* n (factorial (- n 1)))))) (assert-equal 120 (factorial 5)))')
+
+    def test_higher_order_returns_lambda(self):
+        _run('(let ((make-adder (fn (n) (fn (x) (+ n x))))) (let ((add5 (make-adder 5))) (assert-equal 8 (add5 3))))')
+
+
+class TestSpecHigherOrder:
+    """test.sx suite: higher-order"""
+
+    def test_map(self):
+        _run('(do (assert-equal (list 2 4 6) (map (fn (x) (* x 2)) (list 1 2 3))) (assert-equal (list) (map (fn (x) x) (list))))')
+
+    def test_filter(self):
+        _run('(do (assert-equal (list 2 4) (filter (fn (x) (= (mod x 2) 0)) (list 1 2 3 4))) (assert-equal (list) (filter (fn (x) false) (list 1 2 3))))')
+
+    def test_reduce(self):
+        _run('(do (assert-equal 10 (reduce (fn (acc x) (+ acc x)) 0 (list 1 2 3 4))) (assert-equal 0 (reduce (fn (acc x) (+ acc x)) 0 (list))))')
+
+    def test_some(self):
+        _run('(do (assert-true (some (fn (x) (> x 3)) (list 1 2 3 4 5))) (assert-false (some (fn (x) (> x 10)) (list 1 2 3))))')
+
+    def test_every(self):
+        _run('(do (assert-true (every? (fn (x) (> x 0)) (list 1 2 3))) (assert-false (every? (fn (x) (> x 2)) (list 1 2 3))))')
+
+    def test_map_indexed(self):
+        _run('(assert-equal (list "0:a" "1:b" "2:c") (map-indexed (fn (i x) (str i ":" x)) (list "a" "b" "c")))')
+
+
+class TestSpecComponents:
+    """test.sx suite: components"""
+
+    def test_defcomp_creates_component(self):
+        _run('(do (defcomp ~test-comp (&key title) (div title)) (assert-true (not (nil? ~test-comp))))')
+
+    def test_component_renders_with_keyword_args(self):
+        _run('(do (defcomp ~greeting (&key name) (span (str "Hello, " name "!"))) (assert-true (not (nil? ~greeting))))')
+
+    def test_component_with_children(self):
+        _run('(do (defcomp ~box (&key &rest children) (div :class "box" children)) (assert-true (not (nil? ~box))))')
+
+    def test_component_with_default_via_or(self):
+        _run('(do (defcomp ~label (&key text) (span (or text "default"))) (assert-true (not (nil? ~label))))')
+
+
+class TestSpecMacros:
+    """test.sx suite: macros"""
+
+    def test_defmacro_creates_macro(self):
+        _run('(do (defmacro unless (cond &rest body) (quasiquote (if (not (unquote cond)) (do (splice-unquote body))))) (assert-equal "yes" (unless false "yes")) (assert-nil (unless true "no")))')
+
+    def test_quasiquote_and_unquote(self):
+        _run('(let ((x 42)) (assert-equal (list 1 42 3) (quasiquote (1 (unquote x) 3))))')
+
+    def test_splice_unquote(self):
+        _run('(let ((xs (list 2 3 4))) (assert-equal (list 1 2 3 4 5) (quasiquote (1 (splice-unquote xs) 5))))')
+
+
+class TestSpecThreading:
+    """test.sx suite: threading"""
+
+    def test_thread_first(self):
+        _run('(do (assert-equal 8 (-> 5 (+ 1) (+ 2))) (assert-equal "HELLO" (-> "hello" upcase)) (assert-equal "HELLO WORLD" (-> "hello" (str " world") upcase)))')
+
+
+class TestSpecTruthiness:
+    """test.sx suite: truthiness"""
+
+    def test_truthy_values(self):
+        _run('(do (assert-true (if 1 true false)) (assert-true (if "x" true false)) (assert-true (if (list 1) true false)) (assert-true (if true true false)))')
+
+    def test_falsy_values(self):
+        _run('(do (assert-false (if false true false)) (assert-false (if nil true false)))')
+
+
+class TestSpecEdgeCases:
+    """test.sx suite: edge-cases"""
+
+    def test_nested_let_scoping(self):
+        _run('(let ((x 1)) (let ((x 2)) (assert-equal 2 x)))')
+
+    def test_recursive_map(self):
+        _run('(assert-equal (list (list 2 4) (list 6 8)) (map (fn (sub) (map (fn (x) (* x 2)) sub)) (list (list 1 2) (list 3 4))))')
+
+    def test_keyword_as_value(self):
+        _run('(do (assert-equal "class" :class) (assert-equal "id" :id))')
+
+    def test_dict_with_evaluated_values(self):
+        _run('(let ((x 42)) (assert-equal 42 (get {:val x} "val")))')
+
+    def test_nil_propagation(self):
+        _run('(do (assert-nil (get {:a 1} "missing")) (assert-equal "default" (or (get {:a 1} "missing") "default")))')
+
+    def test_empty_operations(self):
+        _run('(do (assert-equal (list) (map (fn (x) x) (list))) (assert-equal (list) (filter (fn (x) true) (list))) (assert-equal 0 (reduce (fn (acc x) (+ acc x)) 0 (list))) (assert-equal 0 (len (list))) (assert-equal "" (str)))')
+
diff --git a/sx/sx/nav-data.sx b/sx/sx/nav-data.sx
index 823a8e0..a9adafd 100644
--- a/sx/sx/nav-data.sx
+++ b/sx/sx/nav-data.sx
@@ -103,7 +103,8 @@
   (dict :label "Continuations" :href "/specs/continuations")
   (dict :label "call/cc" :href "/specs/callcc")
   (dict :label "Deps" :href "/specs/deps")
-  (dict :label "Router" :href "/specs/router")))
+  (dict :label "Router" :href "/specs/router")
+  (dict :label "Testing" :href "/specs/testing")))
 
 (define isomorphism-nav-items (list
   (dict :label "Roadmap" :href "/isomorphism/")
@@ -196,7 +197,10 @@
         :prose "The deps module analyzes component dependency graphs and classifies components as pure or IO-dependent. Phase 1 (bundling): walks component AST bodies to find transitive ~component references, computes the minimal set needed per page, and collects per-page CSS classes from only the used components. Phase 2 (IO detection): scans component ASTs for references to IO primitive names (from boundary.sx declarations — frag, query, service, current-user, highlight, etc.), computes transitive IO refs through the component graph, and caches the result on each component. Components with no transitive IO refs are pure — they can render anywhere without server data. IO-dependent components must expand server-side. The spec provides the classification; each host's async partial evaluator acts on it (expand IO-dependent server-side, serialize pure for client). All functions are pure — each host bootstraps them to native code via --spec-modules deps. Platform functions (component-deps, component-set-deps!, component-css-classes, component-io-refs, component-set-io-refs!, env-components, regex-find-all, scan-css-classes) are implemented natively per target.")
   (dict :slug "router" :filename "router.sx" :title "Router"
         :desc "Client-side route matching — Flask-style pattern parsing, segment matching, route table search."
-        :prose "The router module provides pure functions for matching URL paths against Flask-style route patterns (e.g. /docs/<slug>). Used by client-side routing (Phase 3) to determine if a page can be rendered locally without a server roundtrip. split-path-segments breaks a path into segments, parse-route-pattern converts patterns into typed segment descriptors, match-route-segments tests a path against a parsed pattern returning extracted params, and find-matching-route searches a route table for the first match. No platform interface needed — uses only pure string and list primitives. Bootstrapped via --spec-modules deps,router.")))
+        :prose "The router module provides pure functions for matching URL paths against Flask-style route patterns (e.g. /docs/<slug>). Used by client-side routing (Phase 3) to determine if a page can be rendered locally without a server roundtrip. split-path-segments breaks a path into segments, parse-route-pattern converts patterns into typed segment descriptors, match-route-segments tests a path against a parsed pattern returning extracted params, and find-matching-route searches a route table for the first match. No platform interface needed — uses only pure string and list primitives. Bootstrapped via --spec-modules deps,router.")
+  (dict :slug "testing" :filename "test.sx" :title "Testing"
+        :desc "Self-hosting test framework — SX tests SX. Bootstraps to pytest and Node.js TAP."
+        :prose "The test spec defines a minimal test framework in SX that bootstraps to every host. Tests are written in SX and verify SX semantics — the language tests itself. The framework uses only primitives already in primitives.sx (assert, equal?, type-of, str, list, len) plus assertion helpers defined in SX (assert-equal, assert-true, assert-false, assert-nil, assert-type, assert-length, assert-contains). Two bootstrap compilers read test.sx and emit native test files: bootstrap_test.py produces a pytest module, bootstrap_test_js.py produces a Node.js TAP script. The same 81 tests run on both platforms, verifying cross-host parity.")))
 
 (define all-spec-items (concat core-spec-items (concat adapter-spec-items (concat browser-spec-items (concat extension-spec-items module-spec-items)))))
 
diff --git a/sx/sx/testing.sx b/sx/sx/testing.sx
new file mode 100644
index 0000000..7ed3e4d
--- /dev/null
+++ b/sx/sx/testing.sx
@@ -0,0 +1,234 @@
+;; Testing spec page — SX tests SX.
+
+(defcomp ~spec-testing-content (&key spec-source)
+  (~doc-page :title "Testing"
+    (div :class "space-y-8"
+
+      ;; Intro
+      (div :class "space-y-4"
+        (p :class "text-lg text-stone-600"
+          "SX tests itself. "
+          (code :class "text-violet-700 text-sm" "test.sx")
+          " is a self-executing test spec — it defines "
+          (code :class "text-violet-700 text-sm" "deftest")
+          " and "
+          (code :class "text-violet-700 text-sm" "defsuite")
+          " as macros, writes 81 test cases, and runs them. Any host that provides five platform functions can evaluate the file directly.")
+        (p :class "text-stone-600"
+          "This is not a test "
+          (em "of") " SX — it is a test " (em "in") " SX. The same s-expressions that define how "
+          (code :class "text-violet-700 text-sm" "if")
+          " works are used to verify that "
+          (code :class "text-violet-700 text-sm" "if")
+          " works. No code generation, no intermediate files — the evaluator runs the spec."))
+
+      ;; Live test runner
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "Run in browser")
+        (p :class "text-stone-600"
+          "This page loaded "
+          (code :class "text-violet-700 text-sm" "sx-browser.js")
+          " to render itself. The same evaluator can run "
+          (code :class "text-violet-700 text-sm" "test.sx")
+          " right here — SX testing SX, in your browser:")
+        (div :class "flex items-center gap-4"
+          (button :id "test-btn"
+            :class "px-4 py-2 rounded-md bg-violet-600 text-white font-medium text-sm hover:bg-violet-700 cursor-pointer"
+            :onclick "sxRunTests('test-sx-source','test-output','test-btn')"
+            "Run 81 tests"))
+        (pre :id "test-output"
+          :class "text-sm font-mono bg-stone-900 text-green-400 rounded-lg p-4 overflow-x-auto max-h-96 overflow-y-auto"
+          :style "display:none"
+          "")
+        ;; Hidden: raw test.sx source for the browser runner
+        (textarea :id "test-sx-source" :style "display:none" spec-source)
+        ;; Load the test runner script
+        (script :src (asset-url "/scripts/sx-test-runner.js")))
+
+      ;; How it works
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "Architecture")
+        (p :class "text-stone-600"
+          "The test framework needs five platform functions. Everything else — macros, assertion helpers, test suites — is pure SX:")
+        (div :class "not-prose bg-stone-100 rounded-lg p-5 mx-auto max-w-3xl"
+          (pre :class "text-sm leading-relaxed whitespace-pre-wrap break-words font-mono text-stone-700"
+"test.sx                       Self-executing: macros + helpers + 81 tests
+    |
+    |--- browser                sx-browser.js evaluates test.sx in this page
+    |
+    |--- run.js                 Injects 5 platform fns, evaluates test.sx
+    |       |
+    |       +-> sx-browser.js   JS evaluator (bootstrapped from spec)
+    |
+    |--- run.py                 Injects 5 platform fns, evaluates test.sx
+            |
+            +-> evaluator.py    Python evaluator
+
+Platform functions (5 total — everything else is pure SX):
+    try-call (thunk)          -> {:ok true} | {:ok false :error \"msg\"}
+    report-pass (name)        -> output pass
+    report-fail (name error)  -> output fail
+    push-suite (name)         -> push suite context
+    pop-suite ()              -> pop suite context")))
+
+      ;; Framework
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "The test framework")
+        (p :class "text-stone-600"
+          "The framework defines two macros and nine assertion helpers, all in SX. The macros are the key — they make "
+          (code :class "text-violet-700 text-sm" "defsuite")
+          " and "
+          (code :class "text-violet-700 text-sm" "deftest")
+          " executable forms, not just declarations:")
+        (div :class "rounded-lg border border-stone-200 bg-white p-5 space-y-3"
+          (h3 :class "text-sm font-medium text-stone-500 uppercase tracking-wide" "Macros")
+          (~doc-code :code
+            (highlight "(defmacro deftest (name &rest body)\n  `(let ((result (try-call (fn () ,@body))))\n     (if (get result \"ok\")\n       (report-pass ,name)\n       (report-fail ,name (get result \"error\")))))\n\n(defmacro defsuite (name &rest items)\n  `(do (push-suite ,name)\n       ,@items\n       (pop-suite)))" "lisp")))
+        (p :class "text-stone-600 text-sm"
+          (code :class "text-violet-700 text-sm" "deftest")
+          " wraps the body in a thunk, passes it to "
+          (code :class "text-violet-700 text-sm" "try-call")
+          " (the one platform function that catches errors), then reports pass or fail. "
+          (code :class "text-violet-700 text-sm" "defsuite")
+          " pushes a name onto the context stack, runs its children, and pops.")
+        (div :class "rounded-lg border border-stone-200 bg-white p-5 space-y-3"
+          (h3 :class "text-sm font-medium text-stone-500 uppercase tracking-wide" "Assertion helpers")
+          (~doc-code :code
+            (highlight "(define assert-equal\n  (fn (expected actual)\n    (assert (equal? expected actual)\n      (str \"Expected \" (str expected) \" but got \" (str actual)))))\n\n(define assert-true  (fn (val) (assert val ...)))\n(define assert-false (fn (val) (assert (not val) ...)))\n(define assert-nil   (fn (val) (assert (nil? val) ...)))\n(define assert-type  (fn (expected-type val) ...))\n(define assert-length (fn (expected-len col) ...))\n(define assert-contains (fn (item col) ...))\n(define assert-throws (fn (thunk) ...))" "lisp"))))
+
+      ;; Example tests
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "Example: SX testing SX")
+        (p :class "text-stone-600"
+          "The test suites cover every language feature. Here is the arithmetic suite testing the evaluator's arithmetic primitives:")
+        (div :class "rounded-lg border border-stone-200 bg-white p-5 space-y-3"
+          (h3 :class "text-sm font-medium text-stone-500 uppercase tracking-wide" "From test.sx")
+          (~doc-code :code
+            (highlight "(defsuite \"arithmetic\"\n  (deftest \"addition\"\n    (assert-equal 3 (+ 1 2))\n    (assert-equal 0 (+ 0 0))\n    (assert-equal -1 (+ 1 -2))\n    (assert-equal 10 (+ 1 2 3 4)))\n\n  (deftest \"subtraction\"\n    (assert-equal 1 (- 3 2))\n    (assert-equal -1 (- 2 3)))\n\n  (deftest \"multiplication\"\n    (assert-equal 6 (* 2 3))\n    (assert-equal 0 (* 0 100))\n    (assert-equal 24 (* 1 2 3 4)))\n\n  (deftest \"division\"\n    (assert-equal 2 (/ 6 3))\n    (assert-equal 2.5 (/ 5 2)))\n\n  (deftest \"modulo\"\n    (assert-equal 1 (mod 7 3))\n    (assert-equal 0 (mod 6 3))))" "lisp"))))
+
+      ;; Running tests — JS
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "JavaScript: direct evaluation")
+        (p :class "text-stone-600"
+          (code :class "text-violet-700 text-sm" "sx-browser.js")
+          " evaluates "
+          (code :class "text-violet-700 text-sm" "test.sx")
+          " directly. The runner injects platform functions and calls "
+          (code :class "text-violet-700 text-sm" "Sx.eval")
+          " on each parsed expression:")
+        (div :class "rounded-lg border border-stone-200 bg-white p-5 space-y-3"
+          (h3 :class "text-sm font-medium text-stone-500 uppercase tracking-wide" "run.js")
+          (~doc-code :code
+            (highlight "var Sx = require('./sx-browser.js');\nvar src = fs.readFileSync('test.sx', 'utf8');\n\nvar env = {\n  'try-call': function(thunk) {\n    try {\n      Sx.eval([thunk], env);   // call the SX lambda\n      return { ok: true };\n    } catch(e) {\n      return { ok: false, error: e.message };\n    }\n  },\n  'report-pass': function(name) { console.log('ok - ' + name); },\n  'report-fail': function(name, err) { console.log('not ok - ' + name); },\n  'push-suite': function(n) { stack.push(n); },\n  'pop-suite':  function()  { stack.pop(); },\n};\n\nvar exprs = Sx.parseAll(src);\nfor (var i = 0; i < exprs.length; i++) Sx.eval(exprs[i], env);" "javascript")))
+        (div :class "rounded-lg border border-stone-200 bg-white p-5 space-y-3"
+          (h3 :class "text-sm font-medium text-stone-500 uppercase tracking-wide" "Output")
+          (~doc-code :code
+            (highlight "$ node shared/sx/tests/run.js\nTAP version 13\nok 1 - literals > numbers are numbers\nok 2 - literals > strings are strings\n...\nok 81 - edge-cases > empty operations\n\n# tests 81\n# pass  81" "bash"))))
+
+      ;; Running tests — Python
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "Python: direct evaluation")
+        (p :class "text-stone-600"
+          "Same approach — the Python evaluator runs "
+          (code :class "text-violet-700 text-sm" "test.sx")
+          " directly:")
+        (div :class "rounded-lg border border-stone-200 bg-white p-5 space-y-3"
+          (h3 :class "text-sm font-medium text-stone-500 uppercase tracking-wide" "run.py")
+          (~doc-code :code
+            (highlight "from shared.sx.parser import parse_all\nfrom shared.sx.evaluator import _eval, _trampoline\n\ndef try_call(thunk):\n    try:\n        _trampoline(_eval([thunk], {}))\n        return {'ok': True}\n    except Exception as e:\n        return {'ok': False, 'error': str(e)}\n\nenv = {\n    'try-call': try_call,\n    'report-pass': report_pass,\n    'report-fail': report_fail,\n    'push-suite': push_suite,\n    'pop-suite': pop_suite,\n}\n\nfor expr in parse_all(src):\n    _trampoline(_eval(expr, env))" "python")))
+        (div :class "rounded-lg border border-stone-200 bg-white p-5 space-y-3"
+          (h3 :class "text-sm font-medium text-stone-500 uppercase tracking-wide" "Output")
+          (~doc-code :code
+            (highlight "$ python shared/sx/tests/run.py\nTAP version 13\nok 1 - literals > numbers are numbers\n...\nok 81 - edge-cases > empty operations\n\n# tests 81\n# pass  81" "bash"))))
+
+      ;; What it proves
+      (div :class "rounded-lg border border-blue-200 bg-blue-50 p-5 space-y-3"
+        (h2 :class "text-lg font-semibold text-blue-900" "What this proves")
+        (ol :class "list-decimal list-inside text-blue-800 space-y-2 text-sm"
+          (li "The test spec is " (strong "written in SX") " and " (strong "executed by SX") " — no code generation")
+          (li "The same 81 tests run on " (strong "Python, Node.js, and in the browser") " from the same file")
+          (li "Each host provides only " (strong "5 platform functions") " — everything else is pure SX")
+          (li "Adding a new host means implementing 5 functions, not rewriting tests")
+          (li "Platform divergences (truthiness of 0, [], \"\") are " (strong "documented, not hidden"))
+          (li "The spec is " (strong "executable") " — click the button above to prove it")))
+
+      ;; Test suites
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "All 15 test suites")
+        (div :class "overflow-x-auto rounded border border-stone-200"
+          (table :class "w-full text-left text-sm"
+            (thead (tr :class "border-b border-stone-200 bg-stone-100"
+              (th :class "px-3 py-2 font-medium text-stone-600" "Suite")
+              (th :class "px-3 py-2 font-medium text-stone-600" "Tests")
+              (th :class "px-3 py-2 font-medium text-stone-600" "Covers")))
+            (tbody
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "literals")
+                (td :class "px-3 py-2" "6")
+                (td :class "px-3 py-2 text-stone-700" "number, string, boolean, nil, list, dict type checking"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "arithmetic")
+                (td :class "px-3 py-2" "5")
+                (td :class "px-3 py-2 text-stone-700" "+, -, *, /, mod with edge cases"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "comparison")
+                (td :class "px-3 py-2" "3")
+                (td :class "px-3 py-2 text-stone-700" "=, equal?, <, >, <=, >="))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "strings")
+                (td :class "px-3 py-2" "7")
+                (td :class "px-3 py-2 text-stone-700" "str, string-length, substring, contains?, upcase, trim, split/join"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "lists")
+                (td :class "px-3 py-2" "10")
+                (td :class "px-3 py-2 text-stone-700" "first, rest, nth, last, cons, append, reverse, empty?, contains?, flatten"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "dicts")
+                (td :class "px-3 py-2" "6")
+                (td :class "px-3 py-2 text-stone-700" "literals, get, assoc, dissoc, keys/vals, has-key?, merge"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "predicates")
+                (td :class "px-3 py-2" "7")
+                (td :class "px-3 py-2 text-stone-700" "nil?, number?, string?, list?, dict?, boolean?, not"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "special-forms")
+                (td :class "px-3 py-2" "10")
+                (td :class "px-3 py-2 text-stone-700" "if, when, cond, and, or, let, let (Clojure), do/begin, define, set!"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "lambdas")
+                (td :class "px-3 py-2" "5")
+                (td :class "px-3 py-2 text-stone-700" "basic, closures, as argument, recursion, higher-order returns"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "higher-order")
+                (td :class "px-3 py-2" "6")
+                (td :class "px-3 py-2 text-stone-700" "map, filter, reduce, some, every?, map-indexed"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "components")
+                (td :class "px-3 py-2" "4")
+                (td :class "px-3 py-2 text-stone-700" "defcomp, &key params, &rest children, defaults"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "macros")
+                (td :class "px-3 py-2" "3")
+                (td :class "px-3 py-2 text-stone-700" "defmacro, quasiquote/unquote, splice-unquote"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "threading")
+                (td :class "px-3 py-2" "1")
+                (td :class "px-3 py-2 text-stone-700" "-> thread-first macro"))
+              (tr :class "border-b border-stone-100"
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "truthiness")
+                (td :class "px-3 py-2" "2")
+                (td :class "px-3 py-2 text-stone-700" "truthy/falsy values (platform-universal subset)"))
+              (tr
+                (td :class "px-3 py-2 font-mono text-sm text-violet-700" "edge-cases")
+                (td :class "px-3 py-2" "6")
+                (td :class "px-3 py-2 text-stone-700" "nested scoping, recursive map, keywords, dict eval, nil propagation, empty ops"))))))
+
+      ;; Full source
+      (div :class "space-y-3"
+        (h2 :class "text-2xl font-semibold text-stone-800" "Full specification source")
+        (p :class "text-xs text-stone-400 italic"
+          "The s-expression source below is the canonical test specification. "
+          "Any host that implements the five platform functions can evaluate it directly.")
+        (div :class "not-prose bg-stone-100 rounded-lg p-5 mx-auto max-w-3xl"
+          (pre :class "text-sm leading-relaxed whitespace-pre-wrap break-words"
+            (code (highlight spec-source "sx"))))))))
diff --git a/sx/sxc/pages/docs.sx b/sx/sxc/pages/docs.sx
index 52bd627..d6f33e3 100644
--- a/sx/sxc/pages/docs.sx
+++ b/sx/sxc/pages/docs.sx
@@ -342,6 +342,8 @@
               :filename (get item "filename") :href (str "/specs/" (get item "slug"))
               :source (read-spec-file (get item "filename"))))
         extension-spec-items))
+    "testing" (~spec-testing-content
+      :spec-source (read-spec-file "test.sx"))
     :else (let ((spec (find-spec slug)))
       (if spec
         (~spec-detail-content