Files
breakpilot-compliance/consent-tester/services/banner_text_checker.py
T
Benjamin Admin 9c11b5463c
CI / detect-changes (push) Successful in 11s
CI / dep-audit (push) Has been skipped
CI / sbom-scan (push) Has been skipped
CI / branch-name (push) Has been skipped
CI / guardrail-integrity (push) Has been skipped
CI / secret-scan (push) Has been skipped
CI / go-lint (push) Has been skipped
CI / validate-canonical-controls (push) Successful in 18s
CI / loc-budget (push) Failing after 17s
CI / python-lint (push) Has been skipped
CI / nodejs-lint (push) Has been skipped
CI / nodejs-build (push) Has been skipped
CI / test-go (push) Has been skipped
CI / iace-gt-coverage (push) Has been skipped
CI / test-python-backend (push) Has been skipped
CI / test-python-document-crawler (push) Has been skipped
CI / test-python-dsms-gateway (push) Has been skipped
fix(audit): P98 + P100 — Cookie-Tabellen-Whitespace + Anpassen-Button-Check
P98: HTML-Tabellen-Zellen wurden bei VW-Cookie-Richtlinie ohne Whitespace
verkettet ('smartSignals2UiDsmartSignals2sUiDsmartSignals2CPs...'). Grund:
el.textContent ignoriert Block-Element-Grenzen. Fix: innerText (whitespace-
respecting) statt textContent. Cookie-Namen werden jetzt einzeln erkannt —
VW-Lauf sollte ~100 Cookies statt 1 finden.

P100: Banner-Check fuer 'Anpassen'/'Einstellungen'-Button im Initial-Banner.
VW-Pattern: nur 2 Buttons (Nur technisch notwendige / Alle akzeptieren),
keine granulare Wahl vor Akzeptanz/Ablehnung. Faktische Manipulation
Richtung Pauschal-Akzeptanz. HIGH-Finding nach EDPB 5/2020 §82.
Pattern: anpassen/einstellungen/cookie-einstellungen/manage cookies/
preferences/customize.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-21 15:08:33 +02:00

532 lines
26 KiB
Python

"""
Banner text legal checks — extracted from consent_scanner.py.
11 checks for cookie banner legal compliance:
1. Impressum link accessible (§5 TMG)
2. DSE link in banner (Art. 13 DSGVO)
3. Wrong DSE consent wording (Art. 13 DSGVO)
4. Reject button visible (§25 TDDDG)
5. Pre-ticked checkboxes (Planet49)
6. Dark pattern button size (EDPB 05/2020)
7. Cookie wall (Phase B check)
8. Re-access to settings (Art. 7(3) DSGVO)
9. Third-party DSE link (Art. 13 DSGVO)
10. Dark-pattern language (EDPB 05/2020)
11. Non-modal dismiss = consent (Planet49)
"""
import logging
from services.script_analyzer import Violation
from services.banner_advanced_checks import run_advanced_checks
from services.banner_dom_walkers import (
SHADOW_BANNER_WALKER_JS,
FOOTER_LABELS_WALKER_JS,
)
logger = logging.getLogger(__name__)
async def check_banner_text(page) -> dict:
"""Check cookie banner text for legal issues.
1. Impressum link must be accessible even with banner overlay (§5 TMG)
2. DSE link must be accessible from banner
3. "Zustimmung zur Datenschutzerklärung" is WRONG — DSE is an information
obligation (Art. 13 DSGVO), not something users "agree" to
"""
violations = []
has_impressum = False
has_dse = False
try:
# Get banner text and links
banner_text = ""
banner_links = []
# Try common banner container selectors
for selector in [
"#CybotCookiebotDialog", "#onetrust-banner-sdk", "#didomi-host",
"#usercentrics-root", ".cky-consent-container", "#cmpbox",
'[class*="cookie-banner"]', '[class*="consent-banner"]',
'[class*="cookie-notice"]', '[role="dialog"]',
]:
try:
el = page.locator(selector).first
if await el.count() > 0:
banner_text = (await el.text_content() or "").strip()
# Get links inside banner
links = await el.locator("a[href]").all()
for link in links:
href = await link.get_attribute("href") or ""
text = (await link.text_content() or "").strip()
banner_links.append({"href": href.lower(), "text": text.lower()})
if banner_text:
break
except Exception:
continue
# P28a + P63: Shadow-DOM Web Component CMPs (Mercedes cmm-cookie-banner,
# BMW cookie-consent-banner). Walker pierces shadow tree + extracts
# label-based legal links (wb7-link/button/role=link). See
# banner_dom_walkers.SHADOW_BANNER_WALKER_JS.
if not banner_text or not banner_links:
try:
shadow_data = await page.evaluate(SHADOW_BANNER_WALKER_JS)
if shadow_data and isinstance(shadow_data, dict):
if shadow_data.get("text"):
banner_text = (banner_text + " " + shadow_data["text"]).strip()
if shadow_data.get("links"):
banner_links.extend(shadow_data["links"])
except Exception:
pass
if not banner_text:
return {"violations": violations, "has_impressum": False, "has_dse": False}
banner_lower = banner_text.lower()
# Check 1: Impressum link in or accessible through banner
has_impressum = any(
"impressum" in l["href"] or "impressum" in l["text"] or
"imprint" in l["href"] or "legal notice" in l["text"]
for l in banner_links
)
# Also check if impressum is visible behind/around banner
if not has_impressum:
try:
imp_visible = await page.locator('a[href*="impressum"], a[href*="imprint"]').first
if await imp_visible.count() > 0 and await imp_visible.is_visible():
has_impressum = True
except Exception:
pass
if not has_impressum:
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text="Impressum nicht aus dem Cookie-Banner erreichbar. "
"Bei ueberlagerndem Banner muss ein Impressum-Link im Banner vorhanden sein (§5 TMG).",
legal_ref="§5 TMG, LG Rostock Az. 3 O 22/19",
))
# Check 2: DSE link in banner
has_dse = any(
"datenschutz" in l["href"] or "datenschutz" in l["text"] or
"privacy" in l["href"] or "privacy" in l["text"] or
"dsgvo" in l["href"]
for l in banner_links
)
if not has_dse:
violations.append(Violation(
service="Cookie-Banner",
severity="MEDIUM",
text="Kein Link zur Datenschutzerklaerung im Cookie-Banner. "
"Nutzer sollten vor der Einwilligung die DSE einsehen koennen.",
legal_ref="Art. 13 DSGVO, ErwGr. 42 DSGVO (informierte Einwilligung)",
))
# Check 3: Wrong wording — "Zustimmung zur Datenschutzerklärung"
wrong_dse_consent_patterns = [
"stimme der datenschutz",
"stimme den datenschutz",
"akzeptiere die datenschutz",
"akzeptiere die privacy",
"agree to the privacy policy",
"accept the privacy",
"datenschutzerklaerung zustimmen",
"datenschutzrichtlinie akzeptieren",
"datenschutzrichtlinie zustimmen",
"i agree to the privacy",
"i accept the privacy",
]
for pattern in wrong_dse_consent_patterns:
if pattern in banner_lower:
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text=f"Falsche Formulierung im Banner: 'Zustimmung zur Datenschutzerklaerung'. "
f"Die DSE ist eine Informationspflicht (Art. 13 DSGVO) — man kann sie "
f"nur zur Kenntnis nehmen, nicht 'zustimmen'. "
f"Korrekt: 'Ich habe die Datenschutzinformationen zur Kenntnis genommen'.",
legal_ref="Art. 13 DSGVO, ErwGr. 42 (informierte Einwilligung ≠ Zustimmung zur DSE)",
))
break
# P28b Check 4: Reject mechanism present + explicit-labeled?
# HIGH = no reject mechanism at all
# MEDIUM = reject available but not labeled "Ablehnen"/"Reject"
# (e.g. only "Nur technisch Notwendige" — semantically
# a reject but EDPB 5/2020 + DSK-OH 2024 prefer explicit
# labeling so users recognize it as the reject option)
# P93: EDPB 5/2020 schreibt kein bestimmtes Wort vor — Reject-
# Mechanismus muss gleichwertig zur Annahme sein. BMW nutzt
# "Cookies verbieten", andere Sites "Tracking ablehnen" o.ae. —
# alle rechtlich gleichwertig.
explicit_reject_texts = [
"ablehnen", "reject", "alle ablehnen",
"decline", "alles ablehnen",
"cookies verbieten", "cookies blockieren",
"tracking ablehnen", "tracking verbieten",
"zurueckweisen", "block all", "deny all",
"alle verweigern", "verweigern",
]
implicit_reject_texts = ["nur notwendige", "nur technisch", "nur essenzielle",
"nur essentielle", "notwendige akzeptieren",
"essential only", "only necessary",
"nur erforderliche"]
has_explicit_reject = any(t in banner_lower for t in explicit_reject_texts)
has_implicit_reject = any(t in banner_lower for t in implicit_reject_texts)
if not has_explicit_reject and not has_implicit_reject:
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text="Kein 'Ablehnen'-Mechanismus im Banner erkannt. "
"Die Ablehnung muss ebenso einfach sein wie die Zustimmung.",
legal_ref="§25 Abs. 1 TDDDG, EDPB Guidelines 05/2020 (Consent)",
))
elif not has_explicit_reject and has_implicit_reject:
violations.append(Violation(
service="Cookie-Banner",
severity="MEDIUM",
text="Reject-Moeglichkeit vorhanden ('Nur technisch Notwendige' o.ae.), "
"aber nicht als 'Ablehnen' beschriftet. Nutzer erkennen 'Ablehnen' "
"schneller als sprachlich umschriebene Varianten. "
"Empfehlung: zusaetzlich 'Ablehnen' als Button-Label.",
legal_ref="EDPB 5/2020 (Consent) + DSK-OH 2024 (Telemedien)",
))
# P100: Granular-Wahl-Pruefung — "Anpassen"/"Einstellungen"-Button
# im Initial-Banner. Wenn er FEHLT (VW-Pattern), ist die granulare
# Cookie-Wahl erst nach Akzeptanz/Ablehnung moeglich — faktische
# Manipulation Richtung "Alle akzeptieren". EDPB 5/2020 §82.
granular_button_texts = [
"anpassen", "einstellungen", "cookie-einstellungen",
"cookies verwalten", "manage cookies", "customize",
"weitere optionen", "more options", "settings",
"individuell", "detaillierte einstellungen",
"praeferenzen", "preferences",
]
has_granular_button = any(t in banner_lower for t in granular_button_texts)
if not has_granular_button:
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text="Granulare Cookie-Auswahl im Initial-Banner nicht "
"moeglich (kein 'Anpassen'/'Einstellungen'-Button). "
"Nutzer koennen nur 'Alle akzeptieren' oder 'Nur "
"technisch notwendige' waehlen — Detailwahl pro "
"Kategorie erst nach Akzeptanz/Ablehnung. Das ist "
"faktische Manipulation Richtung Pauschal-Akzeptanz.",
legal_ref="EDPB Guidelines 5/2020 §82 (granular consent), "
"§25 Abs. 1 TDDDG, Art. 4(11) DSGVO (informierte "
"Einwilligung)",
))
# Check 5: Pre-ticked checkboxes (EuGH Planet49)
try:
pre_checked = await page.evaluate("""
() => {
const banner = document.querySelector(
'#CybotCookiebotDialog, #onetrust-banner-sdk, #didomi-host, '
+ '#usercentrics-root, .cky-consent-container, #cmpbox, '
+ '[class*="cookie-banner"], [class*="consent-banner"], [role="dialog"]'
);
if (!banner) return [];
const checked = banner.querySelectorAll(
'input[type="checkbox"]:checked:not([disabled])'
);
return [...checked]
.filter(cb => {
const label = cb.closest('label')?.textContent || cb.getAttribute('aria-label') || '';
const isNecessary = /notwendig|necessary|essential|erforderlich/i.test(label);
return !isNecessary;
})
.map(cb => cb.closest('label')?.textContent?.trim() || cb.id || 'unknown');
}
""")
if pre_checked:
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text=f"Vorausgewaehlte Checkboxen im Banner: {', '.join(pre_checked[:3])}. "
f"Einwilligung muss durch aktive Handlung erfolgen — vorausgefuellte "
f"Checkboxen sind ungueltig.",
legal_ref="Art. 4(11) DSGVO, EuGH C-673/17 (Planet49)",
))
except Exception:
pass
# Check 6: Dark Pattern — button size/prominence comparison
try:
button_info = await page.evaluate("""
() => {
const banner = document.querySelector(
'#CybotCookiebotDialog, #onetrust-banner-sdk, #didomi-host, '
+ '#usercentrics-root, .cky-consent-container, #cmpbox, '
+ '[class*="cookie-banner"], [class*="consent-banner"], [role="dialog"]'
);
if (!banner) return null;
const buttons = [...banner.querySelectorAll('button, a[role="button"], [class*="btn"]')];
return buttons.slice(0, 6).map(b => {
const style = window.getComputedStyle(b);
const rect = b.getBoundingClientRect();
return {
text: b.textContent?.trim()?.substring(0, 40) || '',
width: rect.width,
height: rect.height,
area: rect.width * rect.height,
bgColor: style.backgroundColor,
fontSize: parseFloat(style.fontSize),
visible: rect.width > 0 && rect.height > 0,
};
});
}
""")
if button_info and len(button_info) >= 2:
accept_btn = None
reject_btn = None
accept_kw = ["akzeptieren", "accept", "zustimmen", "agree", "einverstanden", "ok"]
reject_kw = ["ablehnen", "reject", "notwendige", "decline", "nein",
"technisch", "essenzielle", "essential", "erforderliche"]
for btn in button_info:
text_lower = btn["text"].lower()
if any(k in text_lower for k in accept_kw):
accept_btn = btn
elif any(k in text_lower for k in reject_kw):
reject_btn = btn
if accept_btn and reject_btn:
area_ratio = accept_btn["area"] / max(reject_btn["area"], 1)
if area_ratio > 2.5:
violations.append(Violation(
service="Cookie-Banner",
severity="MEDIUM",
text=f"Dark Pattern: 'Akzeptieren'-Button ist {area_ratio:.1f}x groesser als "
f"'Ablehnen'-Button. Beide Optionen muessen gleichwertig dargestellt werden.",
legal_ref="EDPB Guidelines 05/2020, §25 TDDDG, DSK Orientierungshilfe Telemedien",
))
size_ratio = accept_btn["fontSize"] / max(reject_btn["fontSize"], 1)
if size_ratio > 1.5:
violations.append(Violation(
service="Cookie-Banner",
severity="MEDIUM",
text=f"Dark Pattern: Schriftgroesse 'Akzeptieren' ({accept_btn['fontSize']:.0f}px) "
f"vs. 'Ablehnen' ({reject_btn['fontSize']:.0f}px). "
f"Unterschiedliche Schriftgroessen sind ein Dark Pattern.",
legal_ref="EDPB Guidelines 05/2020 (gleichwertige Darstellung)",
))
except Exception:
pass
# Check 7: Cookie Wall — does rejecting block the site?
# (This is checked in Phase B — if after reject the page is not navigable)
# P29 Check 8: Re-access to cookie settings (Art. 7(3) DSGVO).
# Three quality tiers:
# OK = persistent floating cookie icon OR explicit-labeled
# footer link ("Cookie-Einstellungen", "Cookie-Richtlinie",
# "Cookies verwalten", etc.)
# MEDIUM = re-access only via ambiguous label (e.g. "Einstellungen"
# alone — could mean theme/language) OR only via
# cookies.html doc link (not a settings dialog)
# HIGH = no re-access mechanism found at all
try:
has_floating_icon = False
floating_selectors = [
".cky-btn-revisit", "#ot-sdk-btn", "#ot-sdk-btn-floating",
"[class*='ot-floating']", "[class*='cookie-floating']",
"[id*='cookiebot-renew']", "[class*='cmp-floating']",
"[id*='cmplz-cookiebanner-status']", ".uc-cookie-settings-trigger",
"[class*='consent-floating']", "[data-testid*='cookie-revisit']",
]
for sel in floating_selectors:
try:
if await page.locator(sel).count() > 0:
has_floating_icon = True
break
except Exception:
continue
# Footer label inspection — distinguish explicit vs ambiguous
# P64: OEM design-systems (Mercedes wb7-footer, BMW b-footer) don't
# use <footer>. Scan via evaluate() with multiple candidate roots
# including page-bottom region as last-ditch fallback.
footer_labels: list[str] = []
try:
footer_labels = await page.evaluate(FOOTER_LABELS_WALKER_JS) or []
except Exception:
pass
# Explicit, unambiguous cookie/consent labels
explicit_patterns = [
"cookie-einstellungen", "cookie einstellungen",
"cookie-richtlinie", "cookie richtlinie",
"cookie-praeferenzen", "cookie preferences",
"cookies verwalten", "manage cookies",
"datenschutz-einstellungen", "privacy preferences",
"datenschutzeinstellungen", "datenschutz einstellungen",
"cookie consent", "consent settings",
"cookie-banner", "cookies anpassen",
# P64: OEM-typical labels
"tracking-einstellungen", "tracking einstellungen",
"cookie-zustimmung", "consent verwalten",
]
has_explicit_footer = any(
any(p in lbl for p in explicit_patterns)
for lbl in footer_labels
)
# Ambiguous labels — "Einstellungen" alone, generic "Cookies"
ambiguous_patterns = ["einstellungen", "settings", "cookies"]
has_ambiguous_footer = (not has_explicit_footer) and any(
lbl.strip() in ambiguous_patterns
or any(lbl.strip() == p for p in ambiguous_patterns)
for lbl in footer_labels
)
if has_floating_icon or has_explicit_footer:
pass # OK — no violation
elif has_ambiguous_footer:
violations.append(Violation(
service="Cookie-Banner",
severity="MEDIUM",
text="Re-Zugang zu Cookie-Einstellungen nur ueber mehrdeutiges "
"Footer-Label (z.B. 'Einstellungen' oder 'Cookies'). "
"Empfehlung: persistenten Cookie-Icon-Button (Floating) "
"oder explizites Footer-Label 'Cookie-Einstellungen', "
"'Cookie-Richtlinie' o.ae. damit Nutzer den Widerruf "
"ohne Suchen finden.",
legal_ref="Art. 7 Abs. 3 DSGVO + EDPB 5/2020 (informierte, "
"leicht widerrufbare Einwilligung)",
))
else:
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text="Kein erneuter Zugang zu Cookie-Einstellungen gefunden "
"(weder Floating-Icon noch Footer-Link). Widerruf muss "
"ebenso einfach sein wie Erteilung.",
legal_ref="Art. 7 Abs. 3 DSGVO (Widerruf so einfach wie Einwilligung)",
))
except Exception:
pass
# Check 9: Third-party DSE link — consent links to external domain DSE
try:
page_domain = page.url.split("/")[2].replace("www.", "")
for link in banner_links:
href = link["href"]
if not href.startswith("http"):
continue
link_domain = href.split("/")[2].replace("www.", "") if len(href.split("/")) > 2 else ""
if not link_domain:
continue
is_dse_link = any(kw in link["text"] for kw in [
"datenschutz", "privacy", "dsgvo", "data protection",
])
if is_dse_link and link_domain != page_domain:
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text=f"Consent verweist auf Datenschutzerklaerung von {link_domain} "
f"statt auf eigene DSE. Der Verantwortliche muss eine eigene "
f"Datenschutzerklaerung bereitstellen (Art. 13 DSGVO). "
f"Ein Verweis auf die DSE eines Drittanbieters/Auftragsverarbeiters "
f"reicht nicht aus.",
legal_ref="Art. 13 DSGVO (Informationspflichten), Art. 26 DSGVO (gemeinsame Verantwortlichkeit)",
))
break
except Exception:
pass
# Check 10: Dark-Pattern language — "muessen/erforderlich" for non-essential
dark_pattern_phrases = [
("muessen heruntergeladen werden", "heruntergeladen"),
("muessen akzeptiert werden", "akzeptiert"),
("muessen gesetzt werden", "gesetzt"),
("cookies sind erforderlich", "erforderlich"),
("cookies are required", "required"),
("must be downloaded", "downloaded"),
("must be accepted", "accepted"),
("sind zwingend notwendig", "zwingend"),
("unbedingt erforderlich", "unbedingt"),
]
for phrase, keyword in dark_pattern_phrases:
if phrase in banner_lower:
# Check if context is about non-essential cookies
context_essential = any(kw in banner_lower for kw in [
"technisch notwendig", "essential", "strictly necessary",
"unbedingt erforderlich fuer den betrieb",
])
if not context_essential:
violations.append(Violation(
service="Cookie-Banner",
severity="MEDIUM",
text=f"Dark-Pattern-Sprache: '{phrase}' suggeriert technische "
f"Notwendigkeit fuer nicht-essentielle Cookies. Nutzer koennten "
f"den Eindruck gewinnen, eine Zustimmung sei alternativlos.",
legal_ref="EDPB Guidelines 05/2020 Rn. 70, Art. 7(4) DSGVO (freiwillige Einwilligung)",
))
break
# Check 11: Modal dismiss = consent (click outside closes + sets consent)
try:
dismiss_is_consent = await page.evaluate("""
() => {
const dialog = document.querySelector(
'#CybotCookiebotDialog, #onetrust-banner-sdk, #didomi-host, '
+ '#usercentrics-root, .cky-consent-container, #cmpbox, '
+ '[class*="cookie-banner"], [class*="consent-banner"], [role="dialog"]'
);
if (!dialog) return { hasOverlay: false, overlayCloses: false };
// Check for overlay/backdrop elements
const overlays = document.querySelectorAll(
'.overlay, .backdrop, .modal-backdrop, '
+ '[class*="overlay"], [class*="backdrop"], '
+ '[class*="dimmer"], .cdk-overlay-backdrop'
);
let overlayHasClick = false;
for (const ov of overlays) {
const listeners = getEventListeners ? getEventListeners(ov) : {};
if (listeners.click && listeners.click.length > 0) {
overlayHasClick = true;
}
}
// Alternative: check if dialog is non-modal (no inert on background)
const isModal = dialog.getAttribute('aria-modal') === 'true' ||
dialog.hasAttribute('open');
return {
hasOverlay: overlays.length > 0,
overlayHasClick: overlayHasClick,
isModal: isModal,
dialogRole: dialog.getAttribute('role'),
};
}
""")
if dismiss_is_consent and dismiss_is_consent.get("hasOverlay") and not dismiss_is_consent.get("isModal"):
violations.append(Violation(
service="Cookie-Banner",
severity="HIGH",
text="Consent-Dialog ist nicht modal — Klick auf den Hintergrund kann "
"das Fenster schliessen und als Einwilligung gewertet werden. "
"Ein versehentlicher Klick ist keine aktive Einwilligung. "
"Der Dialog muss modal sein (nur explizite Buttons als Optionen).",
legal_ref="EuGH C-673/17 Planet49 (aktive Handlung), Art. 7(1) DSGVO (Nachweispflicht), "
"EDPB Guidelines 05/2020 Rn. 77 (silence/inactivity ≠ consent)",
))
except Exception:
pass
# ── Checks 12-20: Advanced checks ──────────────────────
try:
advanced = await run_advanced_checks(page, banner_text)
violations.extend(advanced)
except Exception as e:
logger.warning("Advanced banner checks failed: %s", e)
except Exception as e:
logger.warning("Banner text check failed: %s", e)
return {"violations": violations, "has_impressum": has_impressum, "has_dse": has_dse}