Harden source info handoff and portal links

Signed-off-by: Yoshifumi Nakamura <nakamura@riken.jp>

Author: yoshifuminakamura

.gitlab-ci.yml

@@ -101,6 +101,7 @@ generate_matrix:
           - "*.md"
           - "docs/**/*"
           - "result_server/**/*"
+          - "requirements-result-server.txt"
           - "config/system_info.csv"
       when: never
     - when: always
@@ -142,6 +143,7 @@ trigger_child_pipeline:
           - "*.md"
           - "docs/**/*"
           - "result_server/**/*"
+          - "requirements-result-server.txt"
           - "config/system_info.csv"
       when: never
     - when: always

docs/ci.md

@@ -230,6 +230,7 @@ Current skip-oriented patterns include:
 - `*.md`
 - `docs/**/*`
 - `result_server/**/*`
+- `requirements-result-server.txt`
 - `config/system_info.csv`
 
 > Synchronization note: this list mirrors the `paths:` entries in

result_server/templates/_filter_dropdowns.html

@@ -19,3 +19,4 @@
             <option value="{{ val }}" {% if current_exp == val %}selected{% endif %}>{{ val }}</option>
         {% endfor %}
     </select></label>
+</div>

result_server/templates/_results_table_cell_code.html

@@ -1,7 +1,7 @@
-{% if row.source_info and row.source_info.source_type == "git" %}
-<td class="code-cell"><a class="code-link" href="{{ row.source_info.repo_url }}" target="_blank" title="{{ row.source_info.repo_url }}">{{ row.code }}</a><span class="quality-badge quality-{{ row.quality.level }}" title="{{ row.quality.label }}: {{ row.quality.summary }}">{{ row.quality.label }}</span></td>
-{% elif row.source_info and row.source_info.source_type == "file" %}
-<td class="code-cell" title="{{ row.source_info.file_path }}"><span class="code-link">{{ row.code }}</span><span class="quality-badge quality-{{ row.quality.level }}" title="{{ row.quality.label }}: {{ row.quality.summary }}">{{ row.quality.label }}</span></td>
+{% if row.source_link and row.source_link.href %}
+<td class="code-cell"><a class="code-link" href="{{ row.source_link.href }}" target="_blank" rel="noopener noreferrer" title="{{ row.source_link.title }}">{{ row.code }}</a><span class="quality-badge quality-{{ row.quality.level }}" title="{{ row.quality.label }}: {{ row.quality.summary }}">{{ row.quality.label }}</span></td>
+{% elif row.source_link %}
+<td class="code-cell" title="{{ row.source_link.title }}"><span class="code-link">{{ row.code }}</span><span class="quality-badge quality-{{ row.quality.level }}" title="{{ row.quality.label }}: {{ row.quality.summary }}">{{ row.quality.label }}</span></td>
 {% else %}
 <td class="code-cell" title="{{ row.code }}"><span class="code-link">{{ row.code }}</span><span class="quality-badge quality-{{ row.quality.level }}" title="{{ row.quality.label }}: {{ row.quality.summary }}">{{ row.quality.label }}</span></td>
 {% endif %}

result_server/tests/test_portal_list_templates.py

@@ -173,6 +173,30 @@ def test_pagination_template_urlencodes_filters_without_inline_javascript():
     assert 'code" onclick="alert(1)' not in html
 
 
+def test_code_cell_adds_noopener_to_source_links():
+    app = build_portal_shell_app(
+        templates_dir=os.path.join(os.path.dirname(__file__), "..", "templates"),
+    )
+    with app.test_request_context("/results"):
+        from flask import render_template
+
+        html = render_template(
+            "_results_table_cell_code.html",
+            row={
+                "code": "qws",
+                "source_link": {
+                    "href": "https://example.invalid/repo.git",
+                    "title": "https://example.invalid/repo.git",
+                },
+                "quality": {"level": "ready", "label": "Ready", "summary": "ok"},
+            },
+        )
+
+    assert 'target="_blank"' in html
+    assert 'rel="noopener noreferrer"' in html
+    assert 'href="https://example.invalid/repo.git"' in html
+
+
 def test_estimated_results_template_renders_table_note():
     app = build_portal_shell_app(
         templates_dir=os.path.join(os.path.dirname(__file__), "..", "templates"),

result_server/tests/test_script_source_info_security.py

@@ -0,0 +1,25 @@
+"""Static checks for source_info handoff scripts."""
+
+from pathlib import Path
+
+
+REPO_ROOT = Path(__file__).resolve().parents[2]
+
+
+def test_result_script_does_not_source_source_info_env():
+    result_script = (REPO_ROOT / "scripts" / "result.sh").read_text(encoding="utf-8")
+
+    assert ". results/source_info.env" not in result_script
+    assert "source results/source_info.env" not in result_script
+    assert "build_source_info_block" in result_script
+    assert "jq -n" in result_script
+
+
+def test_bk_fetch_source_writes_encoded_source_info_values():
+    bk_functions = (REPO_ROOT / "scripts" / "bk_functions.sh").read_text(encoding="utf-8")
+
+    assert "BK_SOURCE_INFO_FORMAT=base64-v1" in bk_functions
+    assert "BK_REPO_URL_B64" in bk_functions
+    assert "BK_FILE_PATH_B64" in bk_functions
+    assert 'export BK_REPO_URL="$BK_REPO_URL"' not in bk_functions
+    assert 'export BK_FILE_PATH="$BK_FILE_PATH"' not in bk_functions

result_server/tests/test_source_info_properties.py

@@ -11,6 +11,7 @@
 sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
 
 from test_support import install_portal_test_stubs
+from utils.result_table_rows import _build_source_link
 
 install_portal_test_stubs()
 
@@ -77,3 +78,51 @@ def test_file_source_info_structure_valid_with_md5sum(self, md5sum):
         assert "file_path" in source_info
         assert "md5sum" in source_info
         assert MD5SUM_PATTERN.match(source_info["md5sum"]) is not None
+
+
+def test_git_source_link_allows_http_urls_only():
+    source_info = {
+        "source_type": "git",
+        "repo_url": "https://github.com/example/repo.git",
+    }
+
+    link = _build_source_link(source_info)
+
+    assert link["href"] == "https://github.com/example/repo.git"
+    assert link["title"] == "https://github.com/example/repo.git"
+
+
+def test_git_source_link_rejects_javascript_urls():
+    source_info = {
+        "source_type": "git",
+        "repo_url": "javascript:alert(1)",
+    }
+
+    link = _build_source_link(source_info)
+
+    assert link["href"] is None
+    assert link["title"] == "Repository URL is not linkable"
+
+
+def test_git_source_link_rejects_ambiguous_urls():
+    source_info = {
+        "source_type": "git",
+        "repo_url": "https://example.invalid\\@evil.invalid/repo.git",
+    }
+
+    link = _build_source_link(source_info)
+
+    assert link["href"] is None
+    assert link["title"] == "Repository URL is not linkable"
+
+
+def test_file_source_link_uses_basename_only():
+    source_info = {
+        "source_type": "file",
+        "file_path": "/sensitive/path/archive.tar.gz",
+    }
+
+    link = _build_source_link(source_info)
+
+    assert link["href"] is None
+    assert link["title"] == "archive.tar.gz"

result_server/utils/result_table_rows.py

@@ -1,3 +1,6 @@
+import os
+from urllib.parse import urlsplit
+
 from flask import url_for
 
 from utils.result_records import (
@@ -13,6 +16,7 @@ def build_result_table_row(json_filename, result_data, padata_filenames):
     matched_padata = _find_matching_padata_archive(json_filename, result_data, padata_filenames)
     pipeline_timing = result_data.get("pipeline_timing", {})
     source_info = result_data.get("source_info")
+    source_link = _build_source_link(source_info)
     profile_data = result_data.get("profile_data")
 
     ci_trigger = result_data.get("ci_trigger", "-") or "-"
@@ -47,6 +51,7 @@ def build_result_table_row(json_filename, result_data, padata_filenames):
         "run_job": result_data.get("run_job", "-") or "-",
         "pipeline_id": pipeline_id,
         "source_info": source_info,
+        "source_link": source_link,
         "source_hash": _format_source_hash(source_info),
         "quality": summarize_result_quality(result_data),
         "profile_data": profile_data,
@@ -97,6 +102,36 @@ def _format_source_hash(source_info):
     return "-"
 
 
+def _build_source_link(source_info):
+    if not isinstance(source_info, dict):
+        return None
+
+    source_type = source_info.get("source_type")
+    if source_type == "git":
+        repo_url = str(source_info.get("repo_url") or "").strip()
+        parsed = urlsplit(repo_url)
+        has_unsafe_chars = any(ch.isspace() for ch in repo_url) or "\\" in repo_url
+        if parsed.scheme in {"http", "https"} and parsed.netloc and not has_unsafe_chars:
+            return {
+                "href": repo_url,
+                "title": repo_url,
+            }
+        return {
+            "href": None,
+            "title": "Repository URL is not linkable",
+        }
+
+    if source_type == "file":
+        file_path = str(source_info.get("file_path") or "").strip()
+        filename = os.path.basename(file_path) if file_path else "source archive"
+        return {
+            "href": None,
+            "title": filename or "source archive",
+        }
+
+    return None
+
+
 def _format_profile_summary(profile_data):
     if not isinstance(profile_data, dict) or not profile_data:
         return "-"

scripts/bk_functions.sh

@@ -1250,6 +1250,43 @@ bk_emit_overlap() {
   bk_emit_section "overlap:${_bk_ovl_sections}" "$_bk_ovl_time" "$_bk_ovl_package" "$_bk_ovl_artifact" --type overlap --members "$_bk_ovl_sections"
 }
 
+bk_base64_encode_value() {
+  if command -v base64 >/dev/null 2>&1; then
+    printf '%s' "$1" | base64 | tr -d '\r\n'
+    return 0
+  fi
+  if command -v openssl >/dev/null 2>&1; then
+    printf '%s' "$1" | openssl base64 -A | tr -d '\r\n'
+    return 0
+  fi
+  echo "bk_base64_encode_value: neither base64 nor openssl found" >&2
+  return 1
+}
+
+bk_write_source_info_env() {
+  _bk_source_type="$1"
+  _bk_repo_url="${2:-}"
+  _bk_branch="${3:-}"
+  _bk_commit_hash="${4:-}"
+  _bk_file_path="${5:-}"
+  _bk_md5sum="${6:-}"
+
+  if ! command -v base64 >/dev/null 2>&1 && ! command -v openssl >/dev/null 2>&1; then
+    echo "bk_write_source_info_env: neither base64 nor openssl found" >&2
+    return 1
+  fi
+
+  {
+    printf 'BK_SOURCE_INFO_FORMAT=base64-v1\n'
+    printf 'BK_SOURCE_TYPE_B64=%s\n' "$(bk_base64_encode_value "$_bk_source_type")"
+    printf 'BK_REPO_URL_B64=%s\n' "$(bk_base64_encode_value "$_bk_repo_url")"
+    printf 'BK_BRANCH_B64=%s\n' "$(bk_base64_encode_value "$_bk_branch")"
+    printf 'BK_COMMIT_HASH_B64=%s\n' "$(bk_base64_encode_value "$_bk_commit_hash")"
+    printf 'BK_FILE_PATH_B64=%s\n' "$(bk_base64_encode_value "$_bk_file_path")"
+    printf 'BK_MD5SUM_B64=%s\n' "$(bk_base64_encode_value "$_bk_md5sum")"
+  } > results/source_info.env
+}
+
 # bk_fetch_source - Fetch source code and collect metadata.
 #
 # Usage:
@@ -1273,7 +1310,7 @@ bk_emit_overlap() {
 #   BK_MD5SUM       - (file) Full 32-char md5sum
 #
 # Side effects:
-#   Writes results/source_info.env in export format
+#   Writes results/source_info.env as data, not executable shell
 #
 # Returns:
 #   0 - success
@@ -1331,13 +1368,7 @@ bk_fetch_source() {
 
     export BK_BRANCH BK_COMMIT_HASH
 
-    # Write results/source_info.env
-    cat > results/source_info.env <<EOF
-export BK_SOURCE_TYPE="git"
-export BK_REPO_URL="$BK_REPO_URL"
-export BK_BRANCH="$BK_BRANCH"
-export BK_COMMIT_HASH="$BK_COMMIT_HASH"
-EOF
+    bk_write_source_info_env "git" "$BK_REPO_URL" "$BK_BRANCH" "$BK_COMMIT_HASH"
 
   else
     # --- File archive path ---
@@ -1380,12 +1411,7 @@ EOF
       fi
     fi
 
-    # Write results/source_info.env
-    cat > results/source_info.env <<EOF
-export BK_SOURCE_TYPE="file"
-export BK_FILE_PATH="$BK_FILE_PATH"
-export BK_MD5SUM="$BK_MD5SUM"
-EOF
+    bk_write_source_info_env "file" "" "" "" "$BK_FILE_PATH" "$BK_MD5SUM"
 
   fi