Top Risks

• THREAT FEED: extension ID independently confirmed malicious by TPCI chrome-mal-ids — ChatGPT search [browser-hijack] (+2.0, capped).
• Privacy policy fetch failed (HTTPError) — cannot verify data handling for 4M users; scored as no policy.
• brand_mention.is_impersonation=true with confirmed_owner=false despite openai.com domain — unverified publisher claiming ChatGPT brand.
• Search engine override silently redirects all user queries to chatgpt.com with tracking ref=ext parameter.
• No developer display name listed; no verified publisher badge despite 4M installs.

Evidence

• search_provider_override manifest chrome_settings_overrides sets chatgpt.com as default search with ref=ext appended to all queries.
• brand_impersonation store brand_mention.is_impersonation=true, confirmed_owner=false; no verified_publisher badge.
• privacy_policy_fetch_failed api privacy_policy_classification.fetched=false (HTTPError); scored at maximum privacy risk per rubric.
• no_verified_publisher store verified_publisher=false, is_featured_by_google=false; developer_name empty despite 4M installs.
• staleness store months_since_update=18; at boundary of 3-6mo and 6-12mo bands — scored +3.5.
• low_rating store Rating 3.3 across large install base; no specific review red flags detected by automated scan.
• no_cve_no_bad_hosts crx cve_findings_raw empty, bad_host_hits empty, code_findings_raw empty, obfuscation_score=0.0.
• openai_domain_established api openai.com resolves, age_days=3844, cert_count=5105 — domain is established and not throwaway.
• threat_feed_membership api Extension ID listed by TPCI chrome-mal-ids — ChatGPT search [browser-hijack]; score 4.63 -> 6.63 (+2.0 cap 2.5).

Permissions Breakdown

• chrome_settings_overrides.search_provider (default) medium Overrides default search engine to chatgpt.com; user search queries redirected to OpenAI.
• content_scripts on https://*.chatgpt.com/* low Scoped narrowly to chatgpt.com only; limited DOM access on one domain.

Pillar Scores

Bookkeeping