Top Risks

• THREAT FEED: extension ID independently confirmed malicious by TPCI chrome-mal-ids — Search Everywhere with Google Bard/Gemini [browser-hijack] + malext.io malicious_extension_sentry (+2.5, capped).
• Brand impersonation: extension name/title falsely implies Google/Gemini/Bard ownership with no verified publisher status.
• Google's generic privacy policy linked — does not scope data collection to this extension; admits data collection and third-party sharing.
• <all_urls> host permission with content_scripts on every page — broad read/write capability for a gmail-developer extension.
• Developer uses free webmail (bardplugin@gmail.com) with no verified business identity; no domain age signal available.

Evidence

• brand_impersonation store brand_mention.is_impersonation=true; brands_mentioned=[google,gemini,bard]; confirmed_owner=false; developer_domain=gmail.com.
• privacy_policy_generic store Policy is Google's account policy; scope_extension=false, data_collection=true, third_party_sharing=true — worst-case privacy tier.
• host_permissions_all_urls manifest host_permissions=[<all_urls>] + content_scripts_matches=[<all_urls>]; full cross-site content access.
• free_webmail_developer store developer_email=bardplugin@gmail.com; no business website; domain_age_ct.queried=false (free webmail).
• install_url_hijack crx install_url_hijack=true; target=https://gemini.google.com/app — redirects user to Google brand page on install.
• js_external_hosts crx 8 external hosts: api.openai.com, be.chatgptbygoogle.com, chat.openai.com, chatgpt.com, chrome.google.com, gemini.google.com, github.com, www.google.com.
• no_csp manifest content_security_policy=null; MV3 default CSP applies but no explicit restriction on connect-src.
• ai_extension_page_content store AI category extension with <all_urls> content scripts — high risk of page-content exfiltration to AI endpoints.
• threat_feed_membership api Extension ID listed by TPCI chrome-mal-ids — Search Everywhere with Google Bard/Gemini [browser-hijack] + malext.io malicious_extension_sentry; score 4.52 -> 7.02 (+2.5 cap 2.5).

Permissions Breakdown

• notifications medium Can push unsolicited notifications to user; medium nuisance/phishing vector.
• storage low Stores local data; low standalone risk.
• sidePanel low Opens side panel UI; low risk in isolation.
• <all_urls> (host_permission) high Content scripts injected on every site; full page read/write capability.

Pillar Scores

Bookkeeping