Netskope Extension Risk

Detail view · rubric v3.6
← Back to catalog

StylerGPT for ChatGPT

hmmijabfpeeiddcjlckllcogkpcaglhg
Risk Score
4.21
Risk Level: Medium
Recommendation: 🟡 MEDIUM RISK — review
Category AI
Installs 10,000
Rating 4.7
Last updated 2026-06-04
Manifest version MV3
CSP present ❌ no
Developer hello@stylergpt.com
Verified publisher ✅ yes
Featured by Google ✅ yes
Privacy policy link
Web Store open ↗

Top Risks

  • No CSP + function_constructor + dynamic script injection + 5x innerHTML sinks — high DOM-XSS and code execution risk.
  • cookies permission on chatgpt.com enables session token access with no CSP to constrain script execution.
  • Brand impersonation: 'ChatGPT' in name, not confirmed OpenAI owner, verified publisher discount limited.
  • AI extension with scripting on chatgpt.com processes all user chat content — high data-sensitivity surface.
  • Privacy policy scoped but no data retention disclosed and third-party sharing silence noted.

Evidence

  • code_quality_function_constructor crx app.js: new Function('return this') constructor found — remote code execution vector.
  • code_quality_script_src_dynamic crx app.js + chatgpt.js: dynamic <script> element creation — script_src_dynamic signal.
  • code_quality_innerhtml_no_csp crx 5 files (inpage-panel, main, onboarding, popup, welcome) contain innerHTML sinks with no CSP present.
  • no_csp_mv3 manifest content_security_policy is null; MV3 default CSP applies but no custom CSP hardening present.
  • brand_impersonation store brand_mention.is_impersonation=true for 'chatgpt'; confirmed_owner=false; verified publisher present.
  • cookies_permission_chatgpt manifest cookies permission + host_permission on chatgpt.com allows session cookie access.
  • privacy_no_retention api Privacy policy fetched, scoped, data_collection=false but retention=false and third_party_silence=true.
  • js_external_hosts crx 7 external JS hosts including fonts.googleapis.com, tailwindcss.com, github.com — 3 countries.

Permissions Breakdown

  • storage low Stores user preferences locally; low risk.
  • activeTab medium Access to current tab on user action; moderate risk.
  • scripting medium Injects scripts into pages; elevated when paired with cookies and no CSP.
  • alarms low Background scheduling; minimal risk.
  • cookies high Can read/write cookies on chatgpt.com — access to session tokens.
  • https://chatgpt.com/* medium Narrowly scoped to ChatGPT; fits stated function.

Pillar Scores

Permissions4.50
Reputation4.50
Network4.50
Webstore4.50
Maintenance0.00
Privacy2.00
Code Quality7.50
CVE Exposure0.00

Bookkeeping

Rubric v3.6
Scored at 2026-06-16 07:40
Listing SHA d46d44865a16…
Force block — not fired
Score recovered no
Elapsed 30.7s