Top Risks

• Privacy policy is Google's generic policy — admits data collection/3rd-party sharing, not scoped to this extension (+10.0 privacy).
• Brand impersonation: mentions ChatGPT and Claude but developer is unverified (m.qayyum@live.com) (+2.0 reputation).
• No developer name listed and free-webmail email; no verified business identity.
• Content scripts on AI platforms (ChatGPT, Claude) can read sensitive conversation content.
• No CSP declared on MV3 extension contacting AI platform hosts.

Evidence

• generic_privacy_policy store Privacy URL is Google's own policy (myaccount.google.com); scope_extension=false, data_collection=true, third_party_sharing=true.
• brand_impersonation store brand_mention.is_impersonation=true; brands=[claude,chatgpt]; developer is unverified individual (live.com email).
• no_developer_name store developer_name is empty string; no verified publisher badge; free-webmail address.
• content_script_ai_platforms manifest Content scripts injected into chat.openai.com, chatgpt.com, claude.ai — can read full AI conversation text.
• no_csp crx content_security_policy=null; csp_present=false on MV3 extension.
• clean_code_scan crx code_findings_raw=[]; obfuscation_score=0.0; 4 JS files scanned, no suspicious patterns found.
• no_cves crx cve_findings_raw=[]; no known-vulnerable bundled libraries detected.
• low_install_count store 761 installs; no operator cluster siblings; wayback shows no ownership change.

Permissions Breakdown

• storage low Local data persistence; no cross-origin exfil risk on its own.
• activeTab low Transient access to current tab only; low blast radius.
• https://chat.openai.com/* medium Host access to ChatGPT; can read/modify AI session content.
• https://chatgpt.com/* medium Duplicate ChatGPT domain; same session-content exposure.
• https://claude.ai/* medium Host access to Claude; can read/modify AI conversation content.

Pillar Scores

Bookkeeping