diff --git a/src/repo_registry/core/models.py b/src/repo_registry/core/models.py
index 299f4e6..99c6cd7 100644
--- a/src/repo_registry/core/models.py
+++ b/src/repo_registry/core/models.py
@@ -36,6 +36,16 @@ class AnalysisRun:
     scanner_version: str
 
 
+@dataclass(frozen=True)
+class ReviewDecision:
+    id: int
+    repository_id: int
+    analysis_run_id: int | None
+    action: str
+    notes: str
+    created_at: str
+
+
 @dataclass(frozen=True)
 class ObservedFact:
     id: int
diff --git a/src/repo_registry/core/service.py b/src/repo_registry/core/service.py
index df70c06..fa794d9 100644
--- a/src/repo_registry/core/service.py
+++ b/src/repo_registry/core/service.py
@@ -10,6 +10,7 @@ from repo_registry.core.models import (
     ObservedFact,
     Repository,
     RepositoryAbilityMap,
+    ReviewDecision,
     ScanSummary,
     SearchResult,
 )
@@ -112,6 +113,13 @@ class RegistryService:
     def list_capabilities(self) -> list[CapabilitySummary]:
         return self.store.list_capabilities()
 
+    def list_review_decisions(
+        self,
+        repository_id: int,
+        analysis_run_id: int | None = None,
+    ) -> list[ReviewDecision]:
+        return self.store.list_review_decisions(repository_id, analysis_run_id)
+
     def list_observed_facts(
         self,
         repository_id: int,
diff --git a/src/repo_registry/storage/sqlite.py b/src/repo_registry/storage/sqlite.py
index 465b7d1..5120a5f 100644
--- a/src/repo_registry/storage/sqlite.py
+++ b/src/repo_registry/storage/sqlite.py
@@ -21,6 +21,7 @@ from repo_registry.core.models import (
     Repository,
     RepositoryAbilityMap,
     RepositorySnapshot,
+    ReviewDecision,
     SearchResult,
     SourceReference,
 )
@@ -929,6 +930,40 @@ class RegistryStore:
             )
             return int(cursor.lastrowid)
 
+    def list_review_decisions(
+        self,
+        repository_id: int,
+        analysis_run_id: int | None = None,
+    ) -> list[ReviewDecision]:
+        self.get_repository(repository_id)
+        params: tuple[int, ...]
+        where = "WHERE repository_id = ?"
+        params = (repository_id,)
+        if analysis_run_id is not None:
+            where += " AND analysis_run_id = ?"
+            params = (repository_id, analysis_run_id)
+        with self.connect() as connection:
+            rows = connection.execute(
+                f"""
+                SELECT id, repository_id, analysis_run_id, action, notes, created_at
+                FROM review_decisions
+                {where}
+                ORDER BY created_at DESC, id DESC
+                """,
+                params,
+            ).fetchall()
+        return [
+            ReviewDecision(
+                id=row["id"],
+                repository_id=row["repository_id"],
+                analysis_run_id=row["analysis_run_id"],
+                action=row["action"],
+                notes=row["notes"],
+                created_at=row["created_at"],
+            )
+            for row in rows
+        ]
+
     def fail_analysis_run(
         self,
         repository_id: int,
diff --git a/src/repo_registry/web_api/app.py b/src/repo_registry/web_api/app.py
index 57383e6..f49563e 100644
--- a/src/repo_registry/web_api/app.py
+++ b/src/repo_registry/web_api/app.py
@@ -344,6 +344,38 @@ def get_analysis_run(
         raise HTTPException(status_code=404, detail=str(exc)) from exc
 
 
+@app.get("/repos/{repository_id}/review-decisions")
+def list_repository_review_decisions(
+    repository_id: int,
+    service: RegistryService = Depends(get_service),
+) -> list[dict[str, object]]:
+    try:
+        return [
+            asdict(decision)
+            for decision in service.list_review_decisions(repository_id)
+        ]
+    except NotFoundError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+
+
+@app.get("/repos/{repository_id}/analysis-runs/{analysis_run_id}/review-decisions")
+def list_analysis_run_review_decisions(
+    repository_id: int,
+    analysis_run_id: int,
+    service: RegistryService = Depends(get_service),
+) -> list[dict[str, object]]:
+    try:
+        return [
+            asdict(decision)
+            for decision in service.list_review_decisions(
+                repository_id,
+                analysis_run_id,
+            )
+        ]
+    except NotFoundError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+
+
 @app.get("/repos/{repository_id}/observed-facts")
 def list_observed_facts(
     repository_id: int,
diff --git a/src/repo_registry/web_ui/views.py b/src/repo_registry/web_ui/views.py
index c45757b..84ecb14 100644
--- a/src/repo_registry/web_ui/views.py
+++ b/src/repo_registry/web_ui/views.py
@@ -267,6 +267,7 @@ def repository_detail(
     repository = service.get_repository(repository_id)
     runs = service.list_analysis_runs(repository_id)
     ability_map = service.ability_map(repository_id)
+    decisions = service.list_review_decisions(repository_id)
     facts = service.list_observed_facts(repository_id)
     languages = sorted({fact.name for fact in facts if fact.kind == "language"})
     frameworks = sorted({fact.name for fact in facts if fact.kind == "framework"})
@@ -307,6 +308,10 @@ def repository_detail(
         {render_ability_map(asdict(ability_map))}
       </section>
     </div>
+    <section class="panel" style="margin-top:18px">
+      <h2>Review Decisions</h2>
+      {render_review_decisions(decisions)}
+    </section>
     """
     return page(repository.name, body)
 
@@ -336,6 +341,7 @@ def analysis_run_detail(
     repository = service.get_repository(repository_id)
     candidate_graph = service.candidate_graph(repository_id, analysis_run_id)
     facts = service.list_observed_facts(repository_id, analysis_run_id)
+    decisions = service.list_review_decisions(repository_id, analysis_run_id)
     fact_rows = "\n".join(
         f"""
         <tr>
@@ -368,6 +374,8 @@ def analysis_run_detail(
           <thead><tr><th>Kind</th><th>Name</th><th>Path</th><th>Value</th></tr></thead>
           <tbody>{fact_rows or '<tr><td colspan="4" class="muted">No observed facts.</td></tr>'}</tbody>
         </table>
+        <h2>Review Decisions</h2>
+        {render_review_decisions(decisions)}
       </section>
     </div>
     """
@@ -743,6 +751,27 @@ def render_repository_facts(languages: list[str], frameworks: list[str]) -> str:
     return f'<p class="actions">{language_pills}{framework_pills}</p>'
 
 
+def render_review_decisions(decisions: list) -> str:
+    if not decisions:
+        return '<p class="muted">No review decisions yet.</p>'
+    rows = "\n".join(
+        f"""
+        <tr>
+          <td>{escape(decision.action)}</td>
+          <td class="source">{escape(decision.created_at)}</td>
+          <td>{escape(decision.notes)}</td>
+        </tr>
+        """
+        for decision in decisions
+    )
+    return f"""
+    <table>
+      <thead><tr><th>Action</th><th>Created</th><th>Notes</th></tr></thead>
+      <tbody>{rows}</tbody>
+    </table>
+    """
+
+
 def render_candidate_ability_actions(
     ability: dict,
     repository_id: int,
diff --git a/tests/test_registry_service.py b/tests/test_registry_service.py
index 275e8aa..c0b405a 100644
--- a/tests/test_registry_service.py
+++ b/tests/test_registry_service.py
@@ -303,6 +303,9 @@ def test_approve_candidate_graph_publishes_ability_map_once(tmp_path):
 
     candidate_graph = service.candidate_graph(repository.id, summary.analysis_run.id)
     assert candidate_graph.abilities[0].status == "approved"
+    decisions = service.list_review_decisions(repository.id, summary.analysis_run.id)
+    assert decisions[0].action == "approve_candidate_graph"
+    assert decisions[0].notes == "Looks good for the first pass."
 
 
 def test_reject_candidate_ability_excludes_it_from_approval(tmp_path):
diff --git a/tests/test_web_api.py b/tests/test_web_api.py
index 13f49ea..6d6202c 100644
--- a/tests/test_web_api.py
+++ b/tests/test_web_api.py
@@ -155,6 +155,17 @@ def test_api_analysis_run_loop(tmp_path):
         )
         assert reject_response.status_code == 200
         assert reject_response.json()["abilities"][0]["status"] == "rejected"
+        decisions_response = client.get(f"/repos/{repository_id}/review-decisions")
+        assert decisions_response.status_code == 200
+        assert decisions_response.json()[0]["action"] == "reject_candidate_ability"
+        run_decisions_response = client.get(
+            f"/repos/{repository_id}/analysis-runs/"
+            f"{run['analysis_run']['id']}/review-decisions"
+        )
+        assert run_decisions_response.status_code == 200
+        assert run_decisions_response.json()[0]["notes"] == (
+            "Reject once to exercise review correction."
+        )
 
         run_response = client.post(f"/repos/{repository_id}/analysis-runs", json={})
         assert run_response.status_code == 201
@@ -302,6 +313,7 @@ def test_ui_register_analyze_and_approve_loop(tmp_path):
         assert run_detail.status_code == 200
         assert "Candidate Graph" in run_detail.text
         assert "ID " in run_detail.text
+        assert "No review decisions yet." in run_detail.text
 
         approve_response = client.post(
             f"{run_path}/candidate-graph/approve",
@@ -316,6 +328,7 @@ def test_ui_register_analyze_and_approve_loop(tmp_path):
         assert "Language: Python" in approved_detail.text
         assert "Framework: FastAPI" in approved_detail.text
         assert "interface:app.py:3" in approved_detail.text
+        assert "approve_candidate_graph" in approved_detail.text
 
         search_response = client.get("/ui/search", params={"q": "repository"})
         assert search_response.status_code == 200