diff --git a/src/skillspector/nodes/analyzers/static_patterns_excessive_agency.py b/src/skillspector/nodes/analyzers/static_patterns_excessive_agency.py
index 04ba47f7..87d56f12 100644
--- a/src/skillspector/nodes/analyzers/static_patterns_excessive_agency.py
+++ b/src/skillspector/nodes/analyzers/static_patterns_excessive_agency.py
@@ -42,7 +42,11 @@
 
 # EA1: Unrestricted Tool Access
 EA1_PATTERNS = [
-    (r"(?:tools?|permissions?)\s*:\s*\[?\s*['\"]?\*['\"]?\s*\]?", 0.85),
+    # EA1 wildcard grant: keep the value on the same line as the colon (the
+    # previous \s* spanned blank lines and bridged unrelated headings) and
+    # require the '*' to be a standalone token so markdown bold (**) or a
+    # word-adjacent '*' (e.g. a YAML anchor) is not treated as a wildcard.
+    (r"(?:tools?|permissions?)\s*:[ \t]*\[?[ \t]*['\"]?\*(?!\*|\w)['\"]?[ \t]*\]?", 0.85),
     (r"(?:allow|grant|enable)\s+(?:access\s+to\s+)?(?:all|any|every)\s+tools?", 0.8),
     (
         r"(?:no|without)\s+(?:tool|permission|access|capability)\s+(?:restrictions?|constraints?|limitations?)",
diff --git a/tests/unit/test_patterns_new.py b/tests/unit/test_patterns_new.py
index d96e5f58..fb7fed8d 100644
--- a/tests/unit/test_patterns_new.py
+++ b/tests/unit/test_patterns_new.py
@@ -88,6 +88,8 @@ class TestExcessiveAgency:
         "content",
         [
             pytest.param('tools: ["*"]', id="wildcard_tools"),
+            pytest.param("tools: *", id="wildcard_tools_bare"),
+            pytest.param("permissions: '*'", id="wildcard_permissions"),
             pytest.param("Allow access to all tools without restriction.", id="grant_all_tools"),
             pytest.param("This skill can execute arbitrary commands.", id="arbitrary_commands"),
             pytest.param(
@@ -104,6 +106,21 @@ def test_ea1_read_only_docs_not_flagged(self) -> None:
         )
         assert not any(f.rule_id == "EA1" for f in findings)
 
+    @pytest.mark.parametrize(
+        "content",
+        [
+            pytest.param(
+                "For each tool:\n\n**Input Schema:**",
+                id="no_cross_heading_bridge",
+            ),
+            pytest.param("**API Coverage vs. Workflow Tools:**", id="no_bold_markdown_collision"),
+            pytest.param("Tools: **Read**, **Write**", id="no_bolded_named_tool_list"),
+        ],
+    )
+    def test_ea1_no_false_positive_on_markdown_bold_or_heading(self, content: str) -> None:
+        findings = ea_mod.analyze(content, "SKILL.md", "markdown")
+        assert not any(f.rule_id == "EA1" for f in findings)
+
     @pytest.mark.parametrize(
         "content,filename,filetype",
         [