AI Assistant for Chrome | SidebarGPT
mejjgaogggabifjfjdbnobinfibaamla
Risk Score
6.84
Risk Level:
High
Recommendation:
🟠 HIGH RISK — review
Top Risks
- THREAT FEED: extension ID independently confirmed malicious by TPCI chrome-mal-ids — Palant affiliate fraud cluster (name unconfirmed) [adware] (+2.0, capped).
- Privacy policy admits data collection and third-party sharing but is not scoped to this extension — scores max privacy risk.
- Developer uses free Gmail address with no verified business identity and no developer name listed.
- Content script injected on <all_urls> gives broad page-content access across every site visited.
- new Function() constructor found in all 4 JS files; MV3 with no CSP leaves dynamic code execution unmitigated.
Evidence
- gmail_developer_no_name store Developer email is kalresanikysa@gmail.com; developer_name is empty — free webmail, no business identity.
- privacy_policy_scope_mismatch api Policy fetched but scope_extension=false, data_collection=true, third_party_sharing=true — generic policy admitting data share.
- host_permissions_all_urls manifest <all_urls> host permission + content_scripts on <all_urls>; AI extension reads every page.
- function_constructor_all_files crx new Function() constructor present in all 4 JS files (background, chat, content-sidebar, options).
- no_csp manifest content_security_policy is null; no CSP declared on MV3 extension with dynamic code patterns.
- js_external_host_github crx js_external_hosts includes github.com — external JS dependency outside developer control.
- verified_publisher_featured store verified_publisher=true and is_featured_by_google=true; discounts applied but capped due to privacy issues.
- ai_page_content_processing store AI extension category with <all_urls> content scripts; page content likely sent to external AI API.
- threat_feed_membership api Extension ID listed by TPCI chrome-mal-ids — Palant affiliate fraud cluster (name unconfirmed) [adware]; score 4.84 -> 6.84 (+2.0 cap 2.5).
Permissions Breakdown
- storage low Stores user preferences/config locally; minimal risk.
- tabs medium Can read tab URLs and titles; moderate info exposure.
- <all_urls> (host_permission) high Content script injected on every site; broad read/write access to all page content.
Pillar Scores
Permissions5.00
Reputation4.50
Network2.00
Webstore3.50
Maintenance3.50
Privacy10.00
Code Quality2.50
CVE Exposure0.00
Bookkeeping
Rubric v3.6
Scored at 2026-11-03 09:52
Listing SHA
0f249622cd8b…
Force block
— not fired
Score recovered
no
Elapsed
—