From 10ec9b05082be38719e7ae5d4847f526ebf9b184 Mon Sep 17 00:00:00 2001
From: Tom Softreck <tom@sapletta.com>
Date: Mon, 12 Oct 2026 20:08:07 +0200
Subject: [PATCH] fix(ticket-004): strip inline code spans and expand ignore
 dirs in sumd

---
 project/TICKETS.md             |  1 +
 project/ticket-004/README.md   | 24 ++++++++++++++++++++++++
 project/ticket-004/intent.json | 25 +++++++++++++++++++++++++
 sumd/extractor.py              |  8 ++++++++
 sumd/validator.py              |  7 +++++--
 5 files changed, 63 insertions(+), 2 deletions(-)
 create mode 100644 project/ticket-004/README.md
 create mode 100644 project/ticket-004/intent.json

diff --git a/project/TICKETS.md b/project/TICKETS.md
index 3516b17..031ff68 100644
--- a/project/TICKETS.md
+++ b/project/TICKETS.md
@@ -5,4 +5,5 @@
 | :--- | :--- | :--- | :--- | :--- | :--- | :--- |
 | **ticket-001** | [`README.md`](./ticket-001/README.md) | - | - | - | - | - |
 | **ticket-002** | [`README.md`](./ticket-002/README.md) | - | - | - | - | - |
+| **ticket-003** | [`README.md`](./ticket-003/README.md) | - | - | - | - | - |
 <!-- AUTO:TICKET_INDEX:END -->
diff --git a/project/ticket-004/README.md b/project/ticket-004/README.md
new file mode 100644
index 0000000..0e59d0e
--- /dev/null
+++ b/project/ticket-004/README.md
@@ -0,0 +1,24 @@
+# Ticket 004: Fix codeblock link false-positive and expand ignore patterns in sumd
+
+- **ID**: ticket-004
+- **Owner**: unresolved:human
+- **Status**: IN_PROGRESS
+- **Workflow state**: EDIT
+- **Created**: 2026-10-12
+
+## Goal and scope
+
+1. Strip tilde code fences (`~~~`) and inline code spans (`` `...` ``) in addition to backtick fences in `_strip_fenced_code_blocks` (`sumd/validator.py`) so code syntax like array indexing or computed properties is not mistaken for empty markdown links.
+2. Add commonly ignored directories to `_IGNORE_DIRS` in `sumd/extractor.py` (`managed_components`, `vendor`, `.pytest_tmp`, `.test-cache`, `.code2llm_cache`, `.playwright-browsers`, `.dev-data`, `.dockuri`).
+
+## Acceptance criteria
+
+- [x] AC-01: Scope is approved by a human owner.
+- [x] AC-02: `_strip_fenced_code_blocks` strips backticks, tildes, and inline code spans.
+- [x] AC-03: `_IGNORE_DIRS` contains common heavy directories and test caches.
+- [x] AC-04: Test suite and governance checks pass.
+
+## Tracking boundary
+
+This directory contains the minimal reviewed intent. Optional participant prose
+and raw command logs are not required delivery output.
diff --git a/project/ticket-004/intent.json b/project/ticket-004/intent.json
new file mode 100644
index 0000000..c5b7077
--- /dev/null
+++ b/project/ticket-004/intent.json
@@ -0,0 +1,25 @@
+{
+  "schema": "new-project.intent/v3",
+  "ticket": "ticket-004",
+  "summary": "Fix codeblock link false-positive and expand ignore patterns in sumd",
+  "workstream": "integration",
+  "classification": {
+    "kind": "SERVICE",
+    "priority": "P2",
+    "origin": "health"
+  },
+  "allowedPaths": [
+    "project/ticket-004/**",
+    "TODO.md",
+    "project/TICKETS.md",
+    "sumd/extractor.py",
+    "sumd/validator.py"
+  ],
+  "forbiddenPaths": [
+    "project/ticket-*/user-*.md"
+  ],
+  "stacks": [],
+  "dependsOn": [],
+  "conflictsWith": [],
+  "integrationTicket": null
+}
diff --git a/sumd/extractor.py b/sumd/extractor.py
index e0e81b9..ccd1151 100644
--- a/sumd/extractor.py
+++ b/sumd/extractor.py
@@ -557,6 +557,14 @@ def extract_package_json(proj_dir: Path) -> dict[str, Any]:
     ".sumd-tools",
     "project",
     "testql-scenarios",
+    "managed_components",
+    "vendor",
+    ".pytest_tmp",
+    ".test-cache",
+    ".code2llm_cache",
+    ".playwright-browsers",
+    ".dev-data",
+    ".dockuri",
 }
 
 # Files to include in M[] (source code only, not docs/config/data)
diff --git a/sumd/validator.py b/sumd/validator.py
index 455e0a6..b6b197d 100644
--- a/sumd/validator.py
+++ b/sumd/validator.py
@@ -290,8 +290,11 @@ def _check_unclosed_fences(lines: list[str], source: str) -> list[str]:
 
 
 def _strip_fenced_code_blocks(content: str) -> str:
-    """Remove fenced code blocks before markdown link validation."""
-    return re.sub(r"```[\s\S]*?```", "", content)
+    """Remove fenced code blocks and inline code spans before markdown link validation."""
+    content = re.sub(r"```[\s\S]*?```", "", content)
+    content = re.sub(r"~~~[\s\S]*?~~~", "", content)
+    content = re.sub(r"`[^`\n]+`", "", content)
+    return content
 
 
 def _check_empty_links(content: str, source: str) -> list[str]: