html2rss 0.29.1 → 0.31.0

data/lib/html2rss/html/feed_link.rb

@@ -2,10 +2,12 @@
 
 module Html2rss
   module Html
+    FeedLink = Data.define(:href, :mime_type)
+
     # A native RSS/Atom hint from a document head +link[rel=alternate]+.
     #
     # Does not guess +/feed+ or +/rss.xml+ paths — that stays in configs +probe_rss+.
-    FeedLink = Data.define(:href, :mime_type) do
+    class FeedLink
       class << self
         ##
         # Collects RSS/Atom +rel=alternate+ links from +doc+'s +head+.

data/lib/html2rss/html/navigator/text_extractor.rb

@@ -19,7 +19,7 @@ module Html2rss
           # @param exclude_nodes [Array<Nokogiri::XML::Node>, nil] nodes to exclude from extraction
           # @return [String, nil] the concatenated visible text, or nil if none is found
           def call(tag, separator: ' ', exclude_nodes: nil)
-            return tag.text.gsub(/\s+/, ' ').strip if tag.respond_to?(:text?) && tag.text?
+            return tag.text.gsub(/\s+/, ' ').strip if tag.is_a?(Nokogiri::XML::Text)
 
             parts = iterate_children(tag, separator, exclude_nodes)
             return if parts.empty?

data/lib/html2rss/html/navigator.rb

@@ -6,16 +6,16 @@ module Html2rss
     # Navigator owns DOM chrome recognition and node traversal helpers.
     class Navigator
       # Heading tags used to prioritize title extraction and container assessment.
-      HEADING_TAGS = %w[h1 h2 h3 h4 h5 h6].freeze
+      HEADING_TAGS = Set['h1', 'h2', 'h3', 'h4', 'h5', 'h6'].freeze
 
       # Element tags that indicate ignored DOM chrome when found in a container path.
-      IGNORED_CONTAINER_TAGS = %w[nav footer header svg script style].to_set.freeze
+      IGNORED_CONTAINER_TAGS = Set['nav', 'footer', 'header', 'svg', 'script', 'style'].freeze
 
       # Ancestor tags that usually indicate navigation/utility regions inside a content container.
-      UTILITY_LANDMARK_TAGS = %w[nav aside footer menu].to_set.freeze
+      UTILITY_LANDMARK_TAGS = Set['nav', 'aside', 'footer', 'menu'].freeze
 
       # Immediate parent walk stops here — not a usable article card.
-      CARD_WALK_STOP_TAGS = (UTILITY_LANDMARK_TAGS | IGNORED_CONTAINER_TAGS | %w[html body]).freeze
+      CARD_WALK_STOP_TAGS = (UTILITY_LANDMARK_TAGS | IGNORED_CONTAINER_TAGS | Set['html', 'body']).freeze
 
       # Inner tags that mean a wrapping <a> is a card, not a span-styled list link.
       WRAPPING_ANCHOR_CHILD_TAGS = (HEADING_TAGS + %w[p]).freeze
@@ -58,7 +58,7 @@ module Html2rss
         # @param condition [Proc] The condition to be met.
         # @return [Nokogiri::XML::Node, nil] The first parent that satisfies the condition.
         def parent_until_condition(node, condition)
-          while node && !node.document? && Probe.tag(node) != 'html'
+          while node && !node.is_a?(Nokogiri::XML::Document) && Probe.tag(node) != 'html'
             return node if condition.call(node)
 
             node = node.parent
@@ -72,7 +72,7 @@ module Html2rss
         # @return [Boolean]
         def usable_card_parent?(node)
           return false unless node
-          return false if node.respond_to?(:document?) && node.document?
+          return false if node.is_a?(Nokogiri::XML::Document)
 
           !CARD_WALK_STOP_TAGS.include?(Probe.tag(node))
         end
@@ -88,8 +88,7 @@ module Html2rss
           while current_tag
             found = current_tag.at_css(selector)
             return found if found
-
-            return nil unless current_tag.respond_to?(:parent)
+            break if current_tag.is_a?(Nokogiri::XML::Document)
 
             current_tag = current_tag.parent
           end
@@ -115,11 +114,12 @@ module Html2rss
         # @param parent_node [Nokogiri::XML::Node] potential ancestor
         # @return [Boolean] true when child_node is a descendant of parent_node
         def descendant_of?(child_node, parent_node)
-          curr = child_node.respond_to?(:parent) ? child_node.parent : nil
+          curr = child_node.is_a?(Nokogiri::XML::Document) ? nil : child_node.parent
           while curr
             return true if curr == parent_node
+            break if curr.is_a?(Nokogiri::XML::Document)
 
-            curr = curr.respond_to?(:parent) ? curr.parent : nil
+            curr = curr.parent
           end
           false
         end

data/lib/html2rss/html/rendering/description_builder.rb

@@ -81,7 +81,7 @@ module Html2rss
 
         def processed_base_description
           text = self.class.remove_pattern_from_start(@base, @title)
-          Html2rss::Selectors::PostProcessors::SanitizeHtml.get(text, @url)
+          Html2rss::Selectors::PostProcessors::SanitizeHtml.call(text, @url)
         end
       end
     end

data/lib/html2rss/html/sst_article_extractor.rb

@@ -12,9 +12,9 @@ module Html2rss
       # CSS class tokens that mark kicker / eyebrow text (excluded from titles).
       KICKER_CLASS_PATTERN = /kicker|eyebrow|pre-title|pretitle|overline/i
       # Inline emphasis tags used as title fallbacks when no heading exists.
-      FALLBACK_HEADING_NAMES = %i[strong b].freeze
+      FALLBACK_HEADING_NAMES = Set[:strong, :b].freeze
       # Nested blocks that mean a category node is actually a content container.
-      CATEGORY_CONTAINER_NAMES = %i[p article section].to_set.freeze
+      CATEGORY_CONTAINER_NAMES = Set[:p, :article, :section].freeze
 
       class << self
         ##
@@ -175,10 +175,13 @@ module Html2rss
       end
 
       def descendant_of?(child, ancestor)
-        # Walk via root children since Extractor may not hold Index; use tag walk on tree.
         return true if child.equal?(ancestor)
 
-        ancestor.descendants.any? { |d| d.equal?(child) }
+        if (index = SST::Index.for_node(child))
+          return index.descendant_of?(child, ancestor)
+        end
+
+        !ancestor.find { |d| d.equal?(child) }.nil?
       end
 
       def leftover_lines
@@ -220,7 +223,7 @@ module Html2rss
       def immediate_card_parent
         parent = sst_parent
         index = SST::Index.for_node(@root)
-        index&.parent_until(parent, method(:usable_walk_parent?))
+        index&.parent_until(parent, ->(node) { usable_walk_parent?(node) })
       end
 
       def usable_walk_parent?(node)

data/lib/html2rss/link_destination/destination_facts.rb

@@ -2,7 +2,6 @@
 
 module Html2rss
   module LinkDestination
-    # Normalized URL plus reusable route-classification facts for one link.
     DestinationFacts = Data.define(
       :url,
       :destination,
@@ -15,7 +14,10 @@ module Html2rss
       :strong_post_suffix,
       :high_confidence_junk_path,
       :high_confidence_utility_destination
-    ) do
+    )
+
+    # Normalized URL plus reusable route-classification facts for one link.
+    class DestinationFacts
       # @param url [Html2rss::Url] normalized destination URL
       # @return [DestinationFacts] route facts for downstream link scoring
       def self.build(url) # rubocop:disable Metrics/MethodLength

data/lib/html2rss/link_destination/noise_policy.rb

@@ -42,7 +42,16 @@ module Html2rss
         destination_facts.utility_path &&
           !destination_facts.content_path &&
           !destination_facts.strong_post_suffix &&
-          text.to_s.scan(/\p{Alnum}+/).size <= 3
+          word_count_at_most?(text, 3)
+      end
+
+      def word_count_at_most?(text, max)
+        count = 0
+        text.to_s.scan(/\p{Alnum}+/) do
+          count += 1
+          return false if count > max
+        end
+        true
       end
 
       def recommended_chrome?(text, destination_facts, heading_anchor:)

data/lib/html2rss/link_destination/path_classifier.rb

@@ -7,47 +7,54 @@ module Html2rss
       attr_reader :segments
 
       # Soft utility segments excluded from high-confidence junk (still utility_path).
-      SOFT_UTILITY = %w[
-        for-you join member members membership newsletter newsletters
-        plans plus premium pricing recommended
-      ].to_set.freeze
+      SOFT_UTILITY = Set[
+        'for-you', 'join', 'member', 'members', 'membership', 'newsletter', 'newsletters',
+        'plans', 'plus', 'premium', 'pricing', 'recommended'
+      ].freeze
 
       # Segment groups used to classify article, taxonomy, utility, and vanity routes.
       # Utility = taxonomy ∪ chrome ∪ soft utility (do not re-list taxonomy tokens).
       SEGMENT_SETS = begin
-        content = %w[
-          article articles blog blogs changelog changelogs insight insights
-          launch launches news post posts release releases story stories update updates
-          artikel beitrag beitraege nachrichten neuigkeiten aktuelles
-          articulo articulos noticia noticias entrada entradas publicacion publicaciones
-          actualite actualites nouvelle nouvelles teaser teasers card cards
-        ].to_set.freeze
-        taxonomy = %w[
-          category categories tag tags topic topics
-          kategorie kategorien schlagwort schlagworte thema themen
-          categoria categorias etiqueta etiquetas tema temas
-          categorie etiquette etiquettes sujet sujets theme themes
-        ].to_set.freeze
-        vanity = %w[
-          join membership plus premium pricing plans subscribe signup
-          abonnieren abo suscribirse boletin s-abonner saboner
-        ].to_set.freeze
-        chrome = %w[
-          about account archive archives author authors comment comments contact feedback
-          help login logout notification notifications preference preferences profile register
-          search settings share signup subscribe feed feeds comment-feed comments-feed privacy
-          terms cookie cookies user users autor autoren archiv ueber-uns ueber ueberuns profil
-          kontakt impressum suche hilfe anmelden registrieren konto registrierung anmeldung
-          abonnieren abo datenschutz nutzungsbedingungen agb autores archivos sobre-nosotros
-          sobre quienes-somos buscar busqueda ayuda entrar ingresar registrarse registro cuenta
-          suscribirse boletin privacidad condiciones auteur auteurs a-propos apropos recherche
-          rechercher aide connexion s-inscrire sinscrire inscription compte s-abonner saboner
-          lettre-information confidentialite mentions-legales cgu menu sidebar widget social
-          modal popup banner promo ad ads related recommendation recommendations pagination pager
-          dating jobs job career careers deals deal shopping shop trading broker versicherung
-          tierversicherung insurance vergleich comparison partnerboerse singleboerse krypto crypto
-          casinos casino kreditkarten kreditkarte kredit echtgeld vpn games kaufberater leasing
-        ].to_set.freeze
+        content = Set[
+          'article', 'articles', 'blog', 'blogs', 'changelog', 'changelogs', 'insight', 'insights',
+          'launch', 'launches', 'news', 'post', 'posts', 'release', 'releases', 'story', 'stories',
+          'update', 'updates', 'artikel', 'beitrag', 'beitraege', 'nachrichten', 'neuigkeiten',
+          'aktuelles', 'articulo', 'articulos', 'noticia', 'noticias', 'entrada', 'entradas',
+          'publicacion', 'publicaciones', 'actualite', 'actualites', 'nouvelle', 'nouvelles',
+          'teaser', 'teasers', 'card', 'cards'
+        ].freeze
+        taxonomy = Set[
+          'category', 'categories', 'tag', 'tags', 'topic', 'topics',
+          'kategorie', 'kategorien', 'schlagwort', 'schlagworte', 'thema', 'themen',
+          'categoria', 'categorias', 'etiqueta', 'etiquetas', 'tema', 'temas',
+          'categorie', 'etiquette', 'etiquettes', 'sujet', 'sujets', 'theme', 'themes'
+        ].freeze
+        vanity = Set[
+          'join', 'membership', 'plus', 'premium', 'pricing', 'plans', 'subscribe', 'signup',
+          'abonnieren', 'abo', 'suscribirse', 'boletin', 's-abonner', 'saboner'
+        ].freeze
+        chrome = Set[
+          'about', 'account', 'archive', 'archives', 'author', 'authors', 'comment', 'comments',
+          'contact', 'feedback', 'help', 'login', 'logout', 'notification', 'notifications',
+          'preference', 'preferences', 'profile', 'register', 'search', 'settings', 'share',
+          'signup', 'subscribe', 'feed', 'feeds', 'comment-feed', 'comments-feed', 'privacy',
+          'terms', 'cookie', 'cookies', 'user', 'users', 'autor', 'autoren', 'archiv', 'ueber-uns',
+          'ueber', 'ueberuns', 'profil', 'kontakt', 'impressum', 'suche', 'hilfe', 'anmelden',
+          'registrieren', 'konto', 'registrierung', 'anmeldung', 'abonnieren', 'abo', 'datenschutz',
+          'nutzungsbedingungen', 'agb', 'autores', 'archivos', 'sobre-nosotros', 'sobre',
+          'quienes-somos', 'buscar', 'busqueda', 'ayuda', 'entrar', 'ingresar', 'registrarse',
+          'registro', 'cuenta', 'suscribirse', 'boletin', 'privacidad', 'condiciones', 'auteur',
+          'auteurs', 'a-propos', 'apropos', 'recherche', 'rechercher', 'aide', 'connexion',
+          's-inscrire', 'sinscrire', 'inscription', 'compte', 's-abonner', 'saboner',
+          'lettre-information', 'confidentialite', 'mentions-legales', 'cgu', 'menu', 'sidebar',
+          'widget', 'social', 'modal', 'popup', 'banner', 'promo', 'ad', 'ads', 'related',
+          'recommendation', 'recommendations', 'pagination', 'pager', 'dating', 'jobs', 'job',
+          'career', 'careers', 'deals', 'deal', 'shopping', 'shop', 'trading', 'broker',
+          'versicherung', 'tierversicherung', 'insurance', 'vergleich', 'comparison',
+          'partnerboerse', 'singleboerse', 'krypto', 'crypto', 'casinos', 'casino',
+          'kreditkarten', 'kreditkarte', 'kredit', 'echtgeld', 'vpn', 'games', 'kaufberater',
+          'leasing'
+        ].freeze
         utility = (taxonomy | chrome | SOFT_UTILITY).freeze
         {
           content:,
@@ -55,7 +62,7 @@ module Html2rss
           high_confidence_junk: (utility - SOFT_UTILITY).freeze,
           taxonomy:,
           vanity:,
-          deep_post_context: %w[press newsroom presse pressemitteilungen prensa].to_set.freeze
+          deep_post_context: Set['press', 'newsroom', 'presse', 'pressemitteilungen', 'prensa'].freeze
         }.freeze
       end
       # Path segment that begins with a year-like publishing marker.

data/lib/html2rss/mcp/README.md

@@ -20,10 +20,12 @@ Batch: `batch_inspect`, `batch_recon`, `batch_scrape`.
 
 ## Decision tree
 
-1. **Articles now (no saved config)?** → `scrape` (or `batch_scrape` for multiple URLs). `strategy: "auto"` runs Faraday → Botasaurus; do not retry with explicit `faraday` after `auto`.
+1. **Articles now (no saved config)?** → `scrape` (or `batch_scrape` for multiple URLs). `strategy: "auto"` runs default (HTTPX) → Botasaurus; do not retry with explicit `default` after `auto`.
 2. **Reusable feed YAML?** → `capture` → `test` → `apply`. `capture` returns YAML in `payload.yaml`. Strive `enhance: true` (false only when chrome leaks). `test` runs schema + live extraction; optional `compare_enhance` compares enhance off vs on. `apply` is the ship gate (`isError` on zero items). Both `test` and `apply` may include `quality_report.enhance_gains` when `selectors.items.enhance` is true.
 3. **Weak scrape/capture or recon?** → `inspect` (or `batch_inspect`). When alternates warrant it, follow `next_step` to `recon`.
 4. **Config already in hand?** → `validate` (schema only) → `test` → `apply`.
+   - `validate` failures return `payload.issues` (array of `{ path, code, message, expected, actual }`) — not a nested `errors` hash.
+   - `test` schema failures surface the same shape under `payload.validation_issues`.
 
 Prefer `capture` for durable configs; `scrape` / `batch_scrape` for one-shot extraction. Follow envelope `next_step` and `guidance`; do not parse scrape text as a raw item array.
 
@@ -61,7 +63,7 @@ Do not duplicate playbook prose in `server.rb`.
 | ----------------------- | ---------------------------------------------------------- |
 | `html2rss://schema`     | Full JSON Schema for feed configurations                   |
 | `html2rss://extractors` | Registered extractor names                                 |
-| `html2rss://strategies` | Published MCP strategies (`auto`, `faraday`, `botasaurus`) |
+| `html2rss://strategies` | Published MCP strategies (`auto`, `default`, `httpx`, `botasaurus`) |
 | `html2rss://runtime`    | `version`, `mcp_contract_version`, `catalog_fingerprint`, `tools`, `botasaurus_configured` (never the scraper URL). Fingerprint covers tool names, required keys, and `oneOf` branches; bump `mcp_contract_version` for envelope semantics. |
 
 ## Prompts
@@ -73,7 +75,7 @@ Do not duplicate playbook prose in `server.rb`.
 
 ## Strategy note
 
-`scrape` / `capture` with `strategy: "auto"` run the full AutoFallback chain. `inspect` maps `auto` to Faraday for cheap diagnostics; pin `botasaurus` when you need browser rendering for inspect.
+`scrape` / `capture` with `strategy: "auto"` run the full AutoFallback chain. `inspect` maps `auto` to default (HTTPX) for cheap diagnostics; pin `botasaurus` when you need browser rendering for inspect.
 
 ## Inspect redirects
 

data/lib/html2rss/mcp/config_argument.rb

@@ -2,10 +2,12 @@
 
 module Html2rss
   module MCP
+    ConfigArgument = Data.define(:config)
+
     # Feed config supplied over MCP as a hash XOR a YAML string.
     #
     # Config does not know MCP. YAML parsing stays on {Html2rss::Config.from_yaml}.
-    ConfigArgument = Data.define(:config) do
+    class ConfigArgument
       class << self
         ##
         # @param config [Hash, nil] feed configuration hash

data/lib/html2rss/mcp/contract.rb

@@ -9,10 +9,14 @@ module Html2rss
     # annotations, and the single compact JSON envelope response.
     module Contract # rubocop:disable Metrics/ModuleLength -- published listing constants stay co-located
       # Published MCP request strategies (excludes +local_file+).
-      STRATEGIES = %w[auto faraday botasaurus].freeze
+      STRATEGIES = %w[auto default httpx botasaurus].freeze
+      # Accepted migration strategies retained for backwards compatibility.
+      MIGRATION_STRATEGIES = %w[faraday].freeze
+      # Complete set of strategies accepted by {.assert_published_request!}.
+      ALL_ACCEPTED_STRATEGIES = (STRATEGIES + MIGRATION_STRATEGIES).freeze
 
       # Bump when tool names, required inputs, or envelope semantics change (independent of gem +VERSION+).
-      MCP_CONTRACT_VERSION = 2
+      MCP_CONTRACT_VERSION = 3
       public_constant :MCP_CONTRACT_VERSION
 
       # Raised when apply/validate config uses an unpublished MCP request adapter.
@@ -30,12 +34,12 @@ module Html2rss
         type: 'string',
         enum: STRATEGIES,
         default: 'auto',
-        description: 'Request strategy (auto runs faraday → botasaurus fallback chain)'
+        description: 'Request strategy (auto runs default → botasaurus fallback chain)'
       }.freeze
 
-      # JSON Schema property for inspect +strategy+ (auto stays on Faraday).
+      # JSON Schema property for inspect +strategy+ (auto stays on default).
       INSPECT_STRATEGY_PROPERTY = STRATEGY_PROPERTY.merge(
-        description: 'Request strategy (auto uses Faraday for cheap diagnostics; pin botasaurus when needed)'
+        description: 'Request strategy (auto uses default for cheap diagnostics; pin botasaurus when needed)'
       ).freeze
 
       # JSON Schema +oneOf+ requiring exactly one of +config+ or +yaml+.
@@ -197,26 +201,12 @@ module Html2rss
       # Tool annotations for +validate+ (closed world).
       ANNOTATIONS_VALIDATE = ANNOTATIONS_OPEN_WORLD.merge(open_world_hint: false).freeze
 
-      # Human titles for +tools/list+.
-      TITLES = {
-        scrape: 'Scrape',
-        inspect: 'Inspect',
-        recon: 'Recon',
-        batch_scrape: 'Batch scrape',
-        batch_inspect: 'Batch inspect',
-        batch_recon: 'Batch recon',
-        capture: 'Capture',
-        validate: 'Validate',
-        apply: 'Apply',
-        test: 'Test'
-      }.freeze
-
       # @api private
       CATALOG_ENTRY_LINE = lambda do |entry|
-        schema = entry.fetch(:input_schema)
+        schema = entry.input_schema
         required = Array(schema[:required]).sort.join(',')
         one_of = Array(schema[:oneOf]).map { |branch| Array(branch[:required]).sort.join('+') }.sort.join('|')
-        [entry.fetch(:name), required, one_of].reject(&:empty?).join(':')
+        [entry.name, required, one_of].reject(&:empty?).join(':')
       end.freeze
 
       class << self
@@ -225,7 +215,7 @@ module Html2rss
         #
         # @return [Array<String>]
         def catalog_tools
-          Server::Tools::TOOLS.map { |entry| entry.fetch(:name) }.sort
+          Server::Tools::TOOLS.map(&:name).sort
         end
 
         ##
@@ -235,7 +225,7 @@ module Html2rss
         #
         # @return [String] 16-char hex digest prefix
         def catalog_fingerprint
-          lines = Server::Tools::TOOLS.sort_by { |entry| entry.fetch(:name) }.map(&CATALOG_ENTRY_LINE)
+          lines = Server::Tools::TOOLS.sort_by(&:name).map(&CATALOG_ENTRY_LINE)
           Digest::SHA256.hexdigest(lines.join("\n")).slice(0, 16)
         end
 
@@ -250,7 +240,7 @@ module Html2rss
             required: %w[ok next_step guidance payload],
             properties: {
               ok: { type: 'boolean' },
-              next_step: { type: 'string', enum: Outcome::NextStep::NAMES.map(&:to_s) },
+              next_step: { type: 'string', enum: Outcome::NextStep::ALL.keys.map(&:to_s) },
               guidance: { type: 'string' },
               payload: { type: 'object' }
             }
@@ -281,7 +271,7 @@ module Html2rss
         #   or +request.local_file_path+ is present
         def assert_published_request!(config)
           strategy = config[:strategy]
-          unless strategy.nil? || STRATEGIES.include?(strategy.to_s)
+          unless strategy.nil? || ALL_ACCEPTED_STRATEGIES.include?(strategy.to_s)
             raise UnpublishedRequestError,
                   "MCP does not accept strategy #{strategy} (published: #{STRATEGIES.join(', ')})"
           end

data/lib/html2rss/mcp/outcome/playbook.rb

@@ -8,7 +8,7 @@ module Html2rss
       # and prompt bodies. {Server} delegates here — do not duplicate prose in
       # +server.rb+.
       class Playbook
-        # Default guidance copy keyed by {Outcome::NextStep::NAMES}.
+        # Default guidance copy; {Outcome::NextStep::ALL} is built from these keys.
         GUIDANCE = {
           done: 'Done. Read payload for the result.',
           inspect: 'Call inspect next. Read payload for diagnostics (final_url, status, ' \
@@ -16,7 +16,7 @@ module Html2rss
           recon: 'Call recon next. Read payload for verdict and native_feed preference.',
           validate: 'Call validate with payload.yaml or a config hash (XOR, not both).',
           apply: 'Call apply next. Confirm payload.item_count before shipping.',
-          scrape: 'Call scrape for articles now. strategy auto already runs Faraday then Botasaurus ' \
+          scrape: 'Call scrape for articles now. strategy auto already runs default (HTTPX) then Botasaurus ' \
                   'and promotes native RSS/Atom when present.',
           capture: 'Call capture for a reusable YAML draft, then follow next_step.',
           read_runtime: 'Read html2rss://runtime. Compare mcp_contract_version and catalog_fingerprint ' \
@@ -36,7 +36,7 @@ module Html2rss
               html2rss MCP — decide which tool to call:
 
               1. Need articles now (no saved config)? → scrape (or batch_scrape for multiple)
-                 - strategy "auto" runs Faraday → Botasaurus AutoFallback. Do not retry with explicit faraday after auto.
+                 - strategy "auto" runs default (HTTPX) → Botasaurus AutoFallback. Do not retry with explicit default after auto.
                  - Empty scrape is still success (articles-now). Follow next_step / guidance (read_runtime if Botasaurus unset).
               2. Need a reusable feed YAML? → capture → test → apply
                  - capture returns YAML inside payload.yaml. Draft only: if destination is html2rss-configs, rewrite for directory.topics and explicit channel title/url. Default enhance follows capture evidence (false when admission_drops show chrome); override only when needed.
@@ -58,9 +58,9 @@ module Html2rss
           # @return [String]
           def scrape_webpage_prompt(url)
             <<~MSG.strip
-              Scrape #{url} with scrape (strategy auto). One call is enough — auto already runs Faraday then Botasaurus.
+              Scrape #{url} with scrape (strategy auto). One call is enough — auto already runs default then Botasaurus.
               Follow envelope next_step and guidance. Call inspect only if articles are empty/weak or you need diagnostics (final_url, status, scheme_downgrade, alternate_feeds). When inspect finds alternates, follow next_step to recon.
-              Do not retry scrape with explicit faraday after auto. Read html2rss://runtime if next_step is read_runtime.
+              Do not retry scrape with explicit default after auto. Read html2rss://runtime if next_step is read_runtime.
               Return payload.items (not a raw JSON array).
             MSG
           end
@@ -72,7 +72,7 @@ module Html2rss
             <<~MSG.strip
               Build a reusable html2rss feed config for #{url}:
               1) capture — YAML is payload.yaml. Check payload.articles_count, payload.has_selectors, and payload.suggested_channel_url. enhance defaults from admission evidence (false when chrome drops are high). When payload.native_feed is set, follow next_step (done — use the native feed).
-              2) Follow next_step. If weak or you need recon, inspect then recon when alternates warrant it. Auto already hops to Botasaurus; do not retry capture with botasaurus unless Faraday was blocked.
+              2) Follow next_step. If weak or you need recon, inspect then recon when alternates warrant it. Auto already hops to Botasaurus; do not retry capture with botasaurus unless default was blocked.
               3) test with yaml (or config hash) — schema + live extraction. On :schema failure, validate; on :execution/:min_items, recapture. Read payload.quality_report.enhance_gains when enhance is on; optional compare_enhance compares enhance off vs on without changing shipped RSS.
               4) apply — isError if zero items. Confirm payload.item_count and payload.quality_report (including enhance_gains) before shipping.
               If the destination is html2rss-configs, rewrite the draft for directory.topics and explicit channel title/url. Return YAML.
@@ -95,7 +95,7 @@ module Html2rss
             if data[:blocked_surface] || data[:surface_category].to_s == 'blocked_surface'
               return 'Blocked or anti-bot interstitial likely. Retry scrape with strategy botasaurus once ' \
                      '(or CLI inspect --deep when BOTASAURUS_SCRAPER_URL is set). ' \
-                     'Do not retry explicit faraday after auto.'
+                     'Do not retry explicit default after auto.'
             end
             if data[:likely_js_shell]
               return 'JS-rendered shell likely (html_present, zero articles). Use strategy auto or botasaurus; ' \