Top Risks

• Privacy policy URL fetch failed — treated as no policy; no scoping, retention, or collection disclosure.
• Brand impersonation: OpenAI mentioned in name/fingerprint by unverified gmail developer.
• Free-webmail developer with no verified business; small install count + high-perm anomaly flagged.
• Content script injected on <all_urls> with innerHTML DOM-XSS sink and no CSP.
• Extension staleness at 17 months with no verified publisher and AI capability processing page content.

Evidence

• privacy_policy_fetch_failed api privacy_policy_classification.fetched=false (fetch_error:HTTPError); scored as no policy → +10.0 privacy.
• brand_impersonation_openai store brand_mention.is_impersonation=true, brands_mentioned=[openai], confirmed_owner=false, dev=gmail.
• free_webmail_dev_no_business store developer_email=stefan.teresiak@gmail.com; developer_domain=gmail.com; no business website.
• dom_xss_sink_no_csp crx dom_sink_innerhtml_userctrl in dist/content.js; csp_present=false amplifies risk to +2.0.
• host_permission_all_urls_content_script manifest host_permissions=[<all_urls>] + content_scripts_matches=[<all_urls>]; broad DOM reach on every site.
• small_install_high_perm_anomaly api install_perm_anomaly.small_install_high_perm=true, tail_attack_surface=true; 70 installs + <all_urls>.
• ai_extension_page_content_processing store AI writing assistant sends page-selected text to api.openai.com; is_ai_extension=true.
• staleness_17_months store months_since_update=17; falls in 12-24mo band → maintenance +6.0.

Permissions Breakdown

• contextMenus low Adds right-click menu items; low standalone risk.
• activeTab medium Accesses current tab on user action; moderate scope.
• storage low Local key/value storage; no exfil risk alone.
• commands low Keyboard shortcut binding; minimal risk.
• <all_urls> (host_permission) high Content script injected on every site; broad DOM access including innerHTML sink.

Pillar Scores

Bookkeeping