diff --git a/lib/parser_inline.js b/lib/parser_inline.js
index e89a71b..b670ba4 100644
--- a/lib/parser_inline.js
+++ b/lib/parser_inline.js
@@ -69,7 +69,7 @@ function ParserInline() {
 // returns `true` if any rule reported success
 //
 ParserInline.prototype.skipToken = function (state) {
-  var i, pos = state.pos,
+  var ok, i, pos = state.pos,
       rules = this.ruler.getRules(''),
       len = rules.length,
       maxNesting = state.md.options.maxNesting,
@@ -81,17 +81,21 @@ ParserInline.prototype.skipToken = function (state) {
     return;
   }
 
-  /*istanbul ignore else*/
   if (state.level < maxNesting) {
     for (i = 0; i < len; i++) {
-      if (rules[i](state, true)) {
-        cache[pos] = state.pos;
-        return;
-      }
+      // Increment state.level and decrement it later to limit recursion.
+      // It's harmless to do here, because no tokens are created. But ideally,
+      // we'd need a separate private state variable for this purpose.
+      //
+      state.level++;
+      ok = rules[i](state, true);
+      state.level--;
+
+      if (ok) { break; }
     }
   }
 
-  state.pos++;
+  if (!ok) { state.pos++; }
   cache[pos] = state.pos;
 };
 
diff --git a/test/misc.js b/test/misc.js
index 1fe0040..a20d495 100644
--- a/test/misc.js
+++ b/test/misc.js
@@ -269,6 +269,22 @@ describe('maxNesting', function () {
     );
   });
 
+  it('Inline parser should not nest above limit', function () {
+    var md = markdownit({ maxNesting: 1 });
+    assert.strictEqual(
+      md.render('[`foo`]()'),
+      '<p><a href="">`foo`</a></p>\n'
+    );
+  });
+
+  it('Inline nesting coverage', function () {
+    var md = markdownit({ maxNesting: 2 });
+    assert.strictEqual(
+      md.render('[[[[[[[[[[[[[[[[[[foo]()'),
+      '<p>[[[[[[[[[[[[[[[[[<a href="">foo</a></p>\n'
+    );
+  });
+
 });