Security
This project is documentation plus small offline scripts, so the realistic risks are credential leaks (your AdGuard Home password) and blocklist changes that break apps on your TV.
Reporting a problem
If you find a security issue in the scripts - for example a way credentials could leak, or a redirect being followed - open a GitHub issue with “[security]” in the title, or reach the maintainer through the repository owner’s GitHub profile. Never include passwords, tokens or your own query logs in a report.
What the scripts do with your data
fetch_query_log.pytalks only to the AdGuard Home address you give it, rejects base URLs containing credentials, queries or fragments before requesting a password, keeps network-error details out of diagnostics, ignores environment/system proxies and connects directly to that address, refuses HTTP redirects (so credentials cannot be forwarded elsewhere), and writes the export atomically to a local file (owner-only mode 0600 on POSIX). A failed write leaves an existing export unchanged. An output symlink is replaced rather than followed; Windows access follows the destination directory ACLs, so use a private directory there. Run it only against your own server.classify_dns_log.pyandvalidate_blocklist.pyread local files and make no network requests at all.- The exported
querylog.jsonis a record of what your TV looked up: treat it as private browsing data, keep it local, and never commit it (it is covered by.gitignore).
Use HTTPS where available: Basic authentication over HTTP is not encrypted.
HTTP remains supported for trusted local installations. Proxy environment variables
such as HTTP_PROXY and HTTPS_PROXY are intentionally ignored by the exporter;
an installation that requires a forward proxy cannot use this tool as-is. A
reverse proxy explicitly included in --host is still supported. URL validation does
not prove ownership of a server; check the address before entering credentials.