From d9f864c0139d14729869ca729f2bc832b3107c50 Mon Sep 17 00:00:00 2001
From: Shirofune-Security
 <43838376+Shirofune-Security@users.noreply.github.com>
Date: Wed, 23 Sep 2026 07:05:56 +0900
Subject: [PATCH] docs: fix issue coverage heading

---
 docs/audit-scoring.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/audit-scoring.md b/docs/audit-scoring.md
index f3c60f5f..3fdc6dbf 100644
--- a/docs/audit-scoring.md
+++ b/docs/audit-scoring.md
@@ -65,6 +65,6 @@ JSON retains the definition/version/hash, profile plan and source provenance, ac
 Version 1.0.0 is defined in `config/audit_scoring.json` as `native-audit-score-v1`. Changing weights requires a reviewed definition-version change, rather than silently moving the denominator. Exact profile and corpus fingerprints let a reviewer identify what was assessed. The profile hash must match before and after planning and the plan's returned hash; observed source changes abort reporting. Hashes bind recorded content, not the trustworthiness of a malicious evidence author. Letter grades and a combined security score are deliberately not defined by this first implementation of issue #10.
 
 Tests cover exact/minimum mask truth tables, optional/role omissions, unknown and empty denominators, severity weights, exclusions/unique IDs, evidence-context preservation, source changes, output collisions, HTML encoding and public command isolation. Windows Server 2022/2025 PowerShell 5.1/7 tests read real policy and verify that native masks/precedence remain unchanged. Synthetic Ready rows test arithmetic only. No Windows 11/DC/AD CS deployment or backend query evidence is claimed by those tests.
-+### Issue 10 coverage
+### Issue 10 coverage
 
 Audit scoring uses weighted rule metadata and reports numerator, denominator, exclusions, and conditional evidence. A score summarizes reviewed eligibility states; it does not prove event generation, forwarding, or detection.