components/reputation/core/safety_tips.proto

// Copyright 2019 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

syntax = "proto2";

option optimize_for = LITE_RUNTIME;

package reputation;

message FlaggedPage {
  enum FlagType {
    UNKNOWN = 0;
    BAD_REP = 1;
    YOUNG_DOMAIN = 2;
  }
  // |pattern| is a full URL, without scheme/username/password/port, such as
  // example.test/test-path-for-safety-tips/test.html.
  optional string pattern = 1;
  optional FlagType type = 2;
}

message UrlPattern {
  // |pattern| is a full URL, without scheme/username/password/port, such as
  // example.test/test-path-for-safety-tips/test.html. Also see the comment for
  // |allowed_pattern| field.
  optional string pattern = 1;
}

message HostPattern {
  // |regex| is a regular expression that matches allowlisted hostnames.
  // This is different than UrlPattern.pattern which matches URLs.
  // IMPORTANT: Don't forget to escape dots when used as label separators.
  optional string regex = 1;
}

// Configuration for the safety tips component. A binary version of this proto
// will be distributed to Chrome clients via component updater. The binary will
// contain a single instance of this message.
message SafetyTipsConfig {
  optional uint32 version_id = 1;

  // List of pages on which to show the Safety Tip UX. This must be sorted and
  // may contain duplicate patterns (when flagged with multiple FlagTypes).
  repeated FlaggedPage flagged_page = 2;

  // List of patterns that are explicitly allowed. This must be sorted. Used to
  // mitigate false positives in Safety Tips and Lookalike warnings.
  // - For safety tips, the pattern can be a URL or a full suffix/prefix
  //   expression used for SafeBrowsing. See
  //   https://developers.google.com/safe-browsing/v4/urls-hashing#suffixprefix-expressions.
  // - Lookalike warnings operate on eTLD+1, so it only makes sense for this
  //   to be eTLD+1, such as "google.com/".
  repeated UrlPattern allowed_pattern = 3;

  // Similar to allowed_pattern, but used to allowlist targets for some of the
  // heuristics.
  // - For edit distance, this is the matched domain. For example, consider edit
  //   distance flagging foo1.com, foo2.com, ... as a spoof of foo.com. If
  //   we are fairly sure that these are all separate and legitimate sites,
  //   allowlisting foo.com is much easier than allowlisting fooN.coms.
  // - For target embedding, this is the embedded target. Some organizations use
  //   lookalike subdomains to proxy popular domains or customize their content
  //   for these popular domains. E.g. google-scholar-com.university.edu.
  //   In these cases it's simpler to allowlist the target instead of the
  //   embedder.
  repeated HostPattern allowed_target_pattern = 4;
}