diff --git a/src/claridoc/lint.py b/src/claridoc/lint.py
index 621160d..8503f51 100644
--- a/src/claridoc/lint.py
+++ b/src/claridoc/lint.py
@@ -15,6 +15,7 @@ from claridoc.models import (
 )
 from claridoc.style_contracts import (
     KOREAN_EXPERIENCE_CONTRACT_ID,
+    first_person_metrics,
     korean_experience_contract_applies,
     plain_form_ending_locations,
 )
@@ -213,6 +214,17 @@ def lint_document(text: str, brief: Brief, outline: Outline, sources: SourcePack
         if style_contract_applies
         else []
     )
+    experience_metrics = (
+        first_person_metrics(text)
+        if style_contract_applies
+        else {
+            "first_person_marker_count": 0,
+            "opening_has_first_person": False,
+            "experience_section_count": 0,
+            "marked_experience_section_count": 0,
+            "experience_section_coverage": 0.0,
+        }
+    )
     if plain_form_locations:
         add(
             "STYLE002",
@@ -227,6 +239,30 @@ def lint_document(text: str, brief: Brief, outline: Outline, sources: SourcePack
                 "current behavior. Preserve quoted material and code unchanged."
             ),
         )
+    if (
+        style_contract_applies
+        and experience_metrics["experience_section_count"]
+        and (
+            not experience_metrics["opening_has_first_person"]
+            or experience_metrics["experience_section_coverage"] < 0.5
+        )
+    ):
+        reasons: list[str] = []
+        if not experience_metrics["opening_has_first_person"]:
+            reasons.append("the opening has no 저는/제가 experience marker")
+        if experience_metrics["experience_section_coverage"] < 0.5:
+            reasons.append(
+                "fewer than half of substantive H2 sections establish first-person experience"
+            )
+        add(
+            "STYLE003",
+            Severity.BLOCKER,
+            "The Korean experience-prose contract is incomplete: " + "; ".join(reasons) + ".",
+            suggestion=(
+                "Use 저는 or 제가 where the opening and major transitions describe "
+                "a supported observation, action, or decision. Do not add invented experience."
+            ),
+        )
     long_paragraph_count = 0
     crowded_paragraph_count = 0
     long_sentence_count = 0
@@ -430,6 +466,7 @@ def lint_document(text: str, brief: Brief, outline: Outline, sources: SourcePack
             else "none"
         ),
         "plain_form_ending_count": len(plain_form_locations),
+        **experience_metrics,
         "has_verification": has_verification,
         "has_tradeoffs": has_tradeoffs,
         "severity_counts": dict(severity_counts),
diff --git a/src/claridoc/style_contracts.py b/src/claridoc/style_contracts.py
index 3bb40b1..b613111 100644
--- a/src/claridoc/style_contracts.py
+++ b/src/claridoc/style_contracts.py
@@ -39,6 +39,7 @@ Write Korean reader-facing prose as a supported first-person experience, not as
 """
 
 _PLAIN_FORM_ENDING = re.compile(r"(?<!니)다(?=[.!?](?:\s|$))")
+_FIRST_PERSON = re.compile(r"(?:저는|제가)")
 _FENCE = re.compile(r"^\s*(?:```|~~~)")
 _HEADING = re.compile(r"^\s{0,3}(#{1,6})\s+(.+?)\s*$")
 _IMAGE_ONLY = re.compile(r"^\s*!\[[^\]]*\]\([^)]*\)\s*$")
@@ -159,3 +160,37 @@ def plain_form_ending_locations(markdown: str) -> list[int]:
     for segment in reader_prose_segments(markdown):
         locations.extend(segment.line for _ in _PLAIN_FORM_ENDING.finditer(segment.text))
     return locations
+
+
+def first_person_metrics(markdown: str) -> dict[str, int | float | bool]:
+    segments = reader_prose_segments(markdown)
+    first_person_marker_count = sum(
+        len(_FIRST_PERSON.findall(segment.text))
+        for segment in segments
+    )
+    opening_has_first_person = bool(
+        segments and _FIRST_PERSON.search(segments[0].text)
+    )
+
+    section_markers: dict[str, bool] = {}
+    for segment in segments:
+        if segment.h2_title is None:
+            continue
+        section_markers.setdefault(segment.h2_title, False)
+        if _FIRST_PERSON.search(segment.text):
+            section_markers[segment.h2_title] = True
+
+    experience_section_count = len(section_markers)
+    marked_experience_section_count = sum(section_markers.values())
+    experience_section_coverage = (
+        marked_experience_section_count / experience_section_count
+        if experience_section_count
+        else 0.0
+    )
+    return {
+        "first_person_marker_count": first_person_marker_count,
+        "opening_has_first_person": opening_has_first_person,
+        "experience_section_count": experience_section_count,
+        "marked_experience_section_count": marked_experience_section_count,
+        "experience_section_coverage": round(experience_section_coverage, 3),
+    }